Nomad changes

packages/elixir/lib/kreuzberg.ex

@@ -0,0 +1,276 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg do
+  @moduledoc "High-level API for kreuzberg"
+
+  @doc "Extract content from a byte array."
+  @spec extract_bytes_async(binary(), String.t()) :: {:ok, map()} | {:error, atom, String.t()}
+  def extract_bytes_async(content, mime_type) do
+    Kreuzberg.Native.extract_bytes_async(content, mime_type, nil)
+  end
+
+  @doc "Extract content from a byte array."
+  @spec extract_bytes_async(binary(), String.t(), String.t() | nil) ::
+          {:ok, map()} | {:error, atom, String.t()}
+  def extract_bytes_async(content, mime_type, config) do
+    Kreuzberg.Native.extract_bytes_async(content, mime_type, config)
+  end
+
+  @doc "Extract content from a file."
+  @spec extract_file_async(String.t(), keyword()) :: {:ok, map()} | {:error, atom, String.t()}
+  def extract_file_async(path, opts \\ []) do
+    Kreuzberg.Native.extract_file_async(
+      path,
+      Keyword.get(opts, :mime_type),
+      Keyword.get(opts, :config)
+    )
+  end
+
+  @doc "Synchronous wrapper for `extract_file`."
+  @spec extract_file_sync(String.t(), keyword()) :: {:ok, map()} | {:error, atom, String.t()}
+  def extract_file_sync(path, opts \\ []) do
+    Kreuzberg.Native.extract_file_sync(
+      path,
+      Keyword.get(opts, :mime_type),
+      Keyword.get(opts, :config)
+    )
+  end
+
+  @doc "Synchronous wrapper for `extract_bytes`."
+  @spec extract_bytes_sync(binary(), String.t()) :: {:ok, map()} | {:error, atom, String.t()}
+  def extract_bytes_sync(content, mime_type) do
+    Kreuzberg.Native.extract_bytes_sync(content, mime_type, nil)
+  end
+
+  @doc "Synchronous wrapper for `extract_bytes`."
+  @spec extract_bytes_sync(binary(), String.t(), String.t() | nil) ::
+          {:ok, map()} | {:error, atom, String.t()}
+  def extract_bytes_sync(content, mime_type, config) do
+    Kreuzberg.Native.extract_bytes_sync(content, mime_type, config)
+  end
+
+  @doc "Detect MIME type from raw file bytes."
+  @spec detect_mime_type_from_bytes(binary()) :: {:ok, String.t()} | {:error, atom, String.t()}
+  def detect_mime_type_from_bytes(content) do
+    Kreuzberg.Native.detect_mime_type_from_bytes(content)
+  end
+
+  @doc "Get file extensions for a given MIME type."
+  @spec get_extensions_for_mime(String.t()) :: {:ok, [String.t()]} | {:error, atom, String.t()}
+  def get_extensions_for_mime(mime_type) do
+    Kreuzberg.Native.get_extensions_for_mime(mime_type)
+  end
+
+  @doc "Clear all embedding backends from the global registry."
+  @spec clear_embedding_backends() :: {:ok, nil} | {:error, atom, String.t()}
+  def clear_embedding_backends do
+    Kreuzberg.Native.clear_embedding_backends()
+  end
+
+  @doc "List the names of all registered embedding backends."
+  @spec list_embedding_backends() :: {:ok, [String.t()]} | {:error, atom, String.t()}
+  def list_embedding_backends do
+    Kreuzberg.Native.list_embedding_backends()
+  end
+
+  @doc "List names of all registered document extractors."
+  @spec list_document_extractors() :: {:ok, [String.t()]} | {:error, atom, String.t()}
+  def list_document_extractors do
+    Kreuzberg.Native.list_document_extractors()
+  end
+
+  @doc "Clear all document extractors from the global registry."
+  @spec clear_document_extractors() :: {:ok, nil} | {:error, atom, String.t()}
+  def clear_document_extractors do
+    Kreuzberg.Native.clear_document_extractors()
+  end
+
+  @doc "List all registered OCR backends."
+  @spec list_ocr_backends() :: {:ok, [String.t()]} | {:error, atom, String.t()}
+  def list_ocr_backends do
+    Kreuzberg.Native.list_ocr_backends()
+  end
+
+  @doc "Clear all OCR backends from the global registry."
+  @spec clear_ocr_backends() :: {:ok, nil} | {:error, atom, String.t()}
+  def clear_ocr_backends do
+    Kreuzberg.Native.clear_ocr_backends()
+  end
+
+  @doc "List all registered post-processor names."
+  @spec list_post_processors() :: {:ok, [String.t()]} | {:error, atom, String.t()}
+  def list_post_processors do
+    Kreuzberg.Native.list_post_processors()
+  end
+
+  @doc "Remove all registered post-processors."
+  @spec clear_post_processors() :: {:ok, nil} | {:error, atom, String.t()}
+  def clear_post_processors do
+    Kreuzberg.Native.clear_post_processors()
+  end
+
+  @doc "List names of all registered renderers."
+  @spec list_renderers() :: {:ok, [String.t()]} | {:error, atom, String.t()}
+  def list_renderers do
+    Kreuzberg.Native.list_renderers()
+  end
+
+  @doc "Clear all renderers from the global registry."
+  @spec clear_renderers() :: {:ok, nil} | {:error, atom, String.t()}
+  def clear_renderers do
+    Kreuzberg.Native.clear_renderers()
+  end
+
+  @doc "List names of all registered validators."
+  @spec list_validators() :: {:ok, [String.t()]} | {:error, atom, String.t()}
+  def list_validators do
+    Kreuzberg.Native.list_validators()
+  end
+
+  @doc "Remove all registered validators."
+  @spec clear_validators() :: {:ok, nil} | {:error, atom, String.t()}
+  def clear_validators do
+    Kreuzberg.Native.clear_validators()
+  end
+
+  @doc "Compare two extraction results and return a structured diff."
+  @spec compare(keyword()) :: map()
+  def compare(opts \\ []) do
+    Kreuzberg.Native.compare(
+      Keyword.get(opts, :a),
+      Keyword.get(opts, :b),
+      Keyword.get(opts, :opts)
+    )
+  end
+
+  @doc "Generate embeddings asynchronously for a list of text strings."
+  @spec embed_texts_async([String.t()]) :: {:ok, [[float()]]} | {:error, atom, String.t()}
+  def embed_texts_async(texts) do
+    Kreuzberg.Native.embed_texts_async(texts, nil)
+  end
+
+  @doc "Generate embeddings asynchronously for a list of text strings."
+  @spec embed_texts_async([String.t()], String.t() | nil) ::
+          {:ok, [[float()]]} | {:error, atom, String.t()}
+  def embed_texts_async(texts, config) do
+    Kreuzberg.Native.embed_texts_async(texts, config)
+  end
+
+  @doc "Render a single PDF page to PNG bytes."
+  @spec render_pdf_page_to_png(binary(), non_neg_integer(), keyword()) ::
+          {:ok, binary()} | {:error, atom, String.t()}
+  def render_pdf_page_to_png(pdf_bytes, page_index, opts \\ []) do
+    Kreuzberg.Native.render_pdf_page_to_png(
+      pdf_bytes,
+      page_index,
+      Keyword.get(opts, :dpi),
+      Keyword.get(opts, :password)
+    )
+  end
+
+  @doc "Detect the MIME type of a file at the given path."
+  @spec detect_mime_type(String.t(), boolean()) :: {:ok, String.t()} | {:error, atom, String.t()}
+  def detect_mime_type(path, check_exists) do
+    Kreuzberg.Native.detect_mime_type(path, check_exists)
+  end
+
+  @doc "Embed a list of texts using the configured embedding model."
+  @spec embed_texts([String.t()]) :: {:ok, [[float()]]} | {:error, atom, String.t()}
+  def embed_texts(texts) do
+    Kreuzberg.Native.embed_texts(texts, nil)
+  end
+
+  @doc "Embed a list of texts using the configured embedding model."
+  @spec embed_texts([String.t()], String.t() | nil) ::
+          {:ok, [[float()]]} | {:error, atom, String.t()}
+  def embed_texts(texts, config) do
+    Kreuzberg.Native.embed_texts(texts, config)
+  end
+
+  @doc "Get an embedding preset by name."
+  @spec get_embedding_preset(String.t()) :: map() | nil
+  def get_embedding_preset(name) do
+    Kreuzberg.Native.get_embedding_preset(name)
+  end
+
+  @doc "List the names of all available embedding presets."
+  @spec list_embedding_presets() :: [String.t()]
+  def list_embedding_presets do
+    Kreuzberg.Native.list_embedding_presets()
+  end
+
+  @doc "Register a OcrBackend plugin with a GenServer PID and name."
+  @spec register_ocr_backend(pid(), String.t()) :: :ok | :error
+  def register_ocr_backend(genserver_pid, plugin_name) do
+    Kreuzberg.Native.register_ocr_backend(genserver_pid, plugin_name)
+  end
+
+  @doc "Unregister a previously registered OcrBackend plugin by name."
+  @spec unregister_ocr_backend(String.t()) :: :ok | :error
+  def unregister_ocr_backend(name) do
+    Kreuzberg.Native.unregister_ocr_backend(name)
+  end
+
+  @doc "Register a PostProcessor plugin with a GenServer PID and name."
+  @spec register_post_processor(pid(), String.t()) :: :ok | :error
+  def register_post_processor(genserver_pid, plugin_name) do
+    Kreuzberg.Native.register_post_processor(genserver_pid, plugin_name)
+  end
+
+  @doc "Unregister a previously registered PostProcessor plugin by name."
+  @spec unregister_post_processor(String.t()) :: :ok | :error
+  def unregister_post_processor(name) do
+    Kreuzberg.Native.unregister_post_processor(name)
+  end
+
+  @doc "Register a Validator plugin with a GenServer PID and name."
+  @spec register_validator(pid(), String.t()) :: :ok | :error
+  def register_validator(genserver_pid, plugin_name) do
+    Kreuzberg.Native.register_validator(genserver_pid, plugin_name)
+  end
+
+  @doc "Unregister a previously registered Validator plugin by name."
+  @spec unregister_validator(String.t()) :: :ok | :error
+  def unregister_validator(name) do
+    Kreuzberg.Native.unregister_validator(name)
+  end
+
+  @doc "Register a EmbeddingBackend plugin with a GenServer PID and name."
+  @spec register_embedding_backend(pid(), String.t()) :: :ok | :error
+  def register_embedding_backend(genserver_pid, plugin_name) do
+    Kreuzberg.Native.register_embedding_backend(genserver_pid, plugin_name)
+  end
+
+  @doc "Unregister a previously registered EmbeddingBackend plugin by name."
+  @spec unregister_embedding_backend(String.t()) :: :ok | :error
+  def unregister_embedding_backend(name) do
+    Kreuzberg.Native.unregister_embedding_backend(name)
+  end
+
+  @doc "Register a DocumentExtractor plugin with a GenServer PID and name."
+  @spec register_document_extractor(pid(), String.t()) :: :ok | :error
+  def register_document_extractor(genserver_pid, plugin_name) do
+    Kreuzberg.Native.register_document_extractor(genserver_pid, plugin_name)
+  end
+
+  @doc "Unregister a previously registered DocumentExtractor plugin by name."
+  @spec unregister_document_extractor(String.t()) :: :ok | :error
+  def unregister_document_extractor(name) do
+    Kreuzberg.Native.unregister_document_extractor(name)
+  end
+
+  @doc "Register a Renderer plugin with a GenServer PID and name."
+  @spec register_renderer(pid(), String.t()) :: :ok | :error
+  def register_renderer(genserver_pid, plugin_name) do
+    Kreuzberg.Native.register_renderer(genserver_pid, plugin_name)
+  end
+
+  @doc "Unregister a previously registered Renderer plugin by name."
+  @spec unregister_renderer(String.t()) :: :ok | :error
+  def unregister_renderer(name) do
+    Kreuzberg.Native.unregister_renderer(name)
+  end
+end

packages/elixir/lib/kreuzberg/acceleration_config.ex

@@ -0,0 +1,48 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.AccelerationConfig do
+  @moduledoc """
+  Hardware acceleration configuration for ONNX Runtime models.
+
+  Controls which execution provider (CPU, CoreML, CUDA, TensorRT) is used
+  for inference in layout detection and embedding generation.
+
+  # Example
+
+  ```rust
+  use kreuzberg::AccelerationConfig;
+
+  // Auto-select: CoreML on macOS, CUDA on Linux, CPU elsewhere
+  let config = AccelerationConfig::default();
+
+  // Force CPU only
+  let config = AccelerationConfig {
+      provider: kreuzberg::ExecutionProviderType::Cpu,
+      ..Default::default()
+  };
+  ```
+  """
+
+  @typedoc "Hardware acceleration configuration for ONNX Runtime models."
+  @type t :: %__MODULE__{
+          provider: String.t() | nil,
+          device_id: non_neg_integer()
+        }
+
+  defstruct provider: :auto,
+            device_id: 0
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/annotation_kind.ex

@@ -0,0 +1,28 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.AnnotationKind do
+  @moduledoc "Types of inline text annotations."
+
+  @typedoc "Types of inline text annotations."
+  @type t :: term()
+
+  @type bold :: :bold
+  @type italic :: :italic
+  @type underline :: :underline
+  @type strikethrough :: :strikethrough
+  @type code :: :code
+  @type subscript :: :subscript
+  @type superscript :: :superscript
+  @type link :: %{type: :link, url: String.t(), title: String.t()}
+  @typedoc "Highlighted text (PDF highlights, HTML `<mark>`)."
+  @type highlight :: :highlight
+  @typedoc "Text color (CSS-compatible value, e.g. \"#ff0000\", \"red\")."
+  @type color :: %{type: :color, value: String.t()}
+  @typedoc "Font size with units (e.g. \"12pt\", \"1.2em\", \"16px\")."
+  @type font_size :: %{type: :font_size, value: String.t()}
+  @typedoc "Extensible annotation for format-specific styling."
+  @type custom :: %{type: :custom, name: String.t(), value: String.t()}
+end

packages/elixir/lib/kreuzberg/archive_entry.ex

@@ -0,0 +1,24 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ArchiveEntry do
+  @moduledoc """
+  A single file extracted from an archive.
+
+  When archives (ZIP, TAR, 7Z, GZIP) are extracted with recursive extraction
+  enabled, each processable file produces its own full `ExtractionResult`.
+  """
+
+  @typedoc "A single file extracted from an archive."
+  @type t :: %__MODULE__{
+          path: String.t() | nil,
+          mime_type: String.t() | nil,
+          result: map()
+        }
+
+  defstruct path: nil,
+            mime_type: nil,
+            result: nil
+end

packages/elixir/lib/kreuzberg/archive_metadata.ex

@@ -0,0 +1,38 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ArchiveMetadata do
+  @moduledoc """
+  Archive (ZIP/TAR/7Z) metadata.
+
+  Extracted from compressed archive files containing file lists and size information.
+  """
+
+  @typedoc "Archive (ZIP/TAR/7Z) metadata."
+  @type t :: %__MODULE__{
+          format: String.t() | nil,
+          file_count: non_neg_integer(),
+          file_list: [String.t()],
+          total_size: non_neg_integer(),
+          compressed_size: non_neg_integer() | nil
+        }
+
+  defstruct format: nil,
+            file_count: 0,
+            file_list: [],
+            total_size: 0,
+            compressed_size: nil
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/b_box.ex

@@ -0,0 +1,21 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.BBox do
+  @moduledoc "Bounding box in original image coordinates (x1, y1) top-left, (x2, y2) bottom-right."
+
+  @typedoc "Bounding box in original image coordinates (x1, y1) top-left, (x2, y2) bottom-right."
+  @type t :: %__MODULE__{
+          x1: float(),
+          y1: float(),
+          x2: float(),
+          y2: float()
+        }
+
+  defstruct x1: 0.0,
+            y1: 0.0,
+            x2: 0.0,
+            y2: 0.0
+end

packages/elixir/lib/kreuzberg/batch_bytes_item.ex

@@ -0,0 +1,24 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.BatchBytesItem do
+  @moduledoc """
+  Batch item for byte array extraction.
+
+  Used with `batch_extract_bytes` and `batch_extract_bytes_sync`
+  to represent a single item in a batch extraction job.
+  """
+
+  @typedoc "Batch item for byte array extraction."
+  @type t :: %__MODULE__{
+          content: binary(),
+          mime_type: String.t() | nil,
+          config: map() | nil
+        }
+
+  defstruct content: <<>>,
+            mime_type: nil,
+            config: nil
+end

packages/elixir/lib/kreuzberg/batch_file_item.ex

@@ -0,0 +1,22 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.BatchFileItem do
+  @moduledoc """
+  Batch item for file extraction.
+
+  Used with `batch_extract_files` and `batch_extract_files_sync`
+  to represent a single file in a batch extraction job.
+  """
+
+  @typedoc "Batch item for file extraction."
+  @type t :: %__MODULE__{
+          path: String.t() | nil,
+          config: map() | nil
+        }
+
+  defstruct path: nil,
+            config: nil
+end

packages/elixir/lib/kreuzberg/bibtex_metadata.ex

@@ -0,0 +1,34 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.BibtexMetadata do
+  @moduledoc "BibTeX bibliography metadata."
+
+  @typedoc "BibTeX bibliography metadata."
+  @type t :: %__MODULE__{
+          entry_count: non_neg_integer(),
+          citation_keys: [String.t()],
+          authors: [String.t()],
+          year_range: map() | nil,
+          entry_types: map() | nil
+        }
+
+  defstruct entry_count: 0,
+            citation_keys: [],
+            authors: [],
+            year_range: nil,
+            entry_types: nil
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/block_type.ex

@@ -0,0 +1,77 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.BlockType do
+  @moduledoc "Types of block-level elements in Djot."
+
+  @typedoc "Types of block-level elements in Djot."
+  @type t ::
+          :paragraph
+          | :heading
+          | :blockquote
+          | :code_block
+          | :list_item
+          | :ordered_list
+          | :bullet_list
+          | :task_list
+          | :definition_list
+          | :definition_term
+          | :definition_description
+          | :div
+          | :section
+          | :thematic_break
+          | :raw_block
+          | :math_display
+
+  @paragraph :paragraph
+  @heading :heading
+  @blockquote :blockquote
+  @code_block :code_block
+  @list_item :list_item
+  @ordered_list :ordered_list
+  @bullet_list :bullet_list
+  @task_list :task_list
+  @definition_list :definition_list
+  @definition_term :definition_term
+  @definition_description :definition_description
+  @div :div
+  @section :section
+  @thematic_break :thematic_break
+  @raw_block :raw_block
+  @math_display :math_display
+
+  @spec paragraph() :: t()
+  def paragraph, do: @paragraph
+  @spec heading() :: t()
+  def heading, do: @heading
+  @spec blockquote() :: t()
+  def blockquote, do: @blockquote
+  @spec code_block() :: t()
+  def code_block, do: @code_block
+  @spec list_item() :: t()
+  def list_item, do: @list_item
+  @spec ordered_list() :: t()
+  def ordered_list, do: @ordered_list
+  @spec bullet_list() :: t()
+  def bullet_list, do: @bullet_list
+  @spec task_list() :: t()
+  def task_list, do: @task_list
+  @spec definition_list() :: t()
+  def definition_list, do: @definition_list
+  @spec definition_term() :: t()
+  def definition_term, do: @definition_term
+  @spec definition_description() :: t()
+  def definition_description, do: @definition_description
+  @spec div() :: t()
+  def div, do: @div
+  @spec section() :: t()
+  def section, do: @section
+  @spec thematic_break() :: t()
+  def thematic_break, do: @thematic_break
+  @spec raw_block() :: t()
+  def raw_block, do: @raw_block
+  @spec math_display() :: t()
+  def math_display, do: @math_display
+end

packages/elixir/lib/kreuzberg/bounding_box.ex

@@ -0,0 +1,32 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.BoundingBox do
+  @moduledoc "Bounding box coordinates for element positioning."
+
+  @typedoc "Bounding box coordinates for element positioning."
+  @type t :: %__MODULE__{
+          x0: float(),
+          y0: float(),
+          x1: float(),
+          y1: float()
+        }
+
+  defstruct x0: 0.0,
+            y0: 0.0,
+            x1: 0.0,
+            y1: 0.0
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/cache_stats.ex

@@ -0,0 +1,22 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.CacheStats do
+  @moduledoc false
+
+  @type t :: %__MODULE__{
+          total_files: non_neg_integer(),
+          total_size_mb: float(),
+          available_space_mb: float(),
+          oldest_file_age_days: float(),
+          newest_file_age_days: float()
+        }
+
+  defstruct total_files: 0,
+            total_size_mb: 0.0,
+            available_space_mb: 0.0,
+            oldest_file_age_days: 0.0,
+            newest_file_age_days: 0.0
+end

packages/elixir/lib/kreuzberg/cell_change.ex

@@ -0,0 +1,27 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.CellChange do
+  @moduledoc """
+  A single changed cell within a table.
+
+  Defined here (rather than only in `crate::diff`) so `RevisionDelta` can
+  reference it unconditionally, without requiring the `diff` Cargo feature.
+  `crate::diff` re-exports this type verbatim.
+  """
+
+  @typedoc "A single changed cell within a table."
+  @type t :: %__MODULE__{
+          row: non_neg_integer(),
+          col: non_neg_integer(),
+          from: String.t() | nil,
+          to: String.t() | nil
+        }
+
+  defstruct row: 0,
+            col: 0,
+            from: nil,
+            to: nil
+end

packages/elixir/lib/kreuzberg/chunk.ex

@@ -0,0 +1,27 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.Chunk do
+  @moduledoc """
+  A text chunk with optional embedding and metadata.
+
+  Chunks are created when chunking is enabled in `ExtractionConfig`. Each chunk
+  contains the text content, optional embedding vector (if embedding generation
+  is configured), and metadata about its position in the document.
+  """
+
+  @typedoc "A text chunk with optional embedding and metadata."
+  @type t :: %__MODULE__{
+          content: String.t() | nil,
+          chunk_type: String.t() | nil,
+          embedding: [float()] | nil,
+          metadata: map()
+        }
+
+  defstruct content: nil,
+            chunk_type: :unknown,
+            embedding: nil,
+            metadata: nil
+end

packages/elixir/lib/kreuzberg/chunk_metadata.ex

@@ -0,0 +1,31 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ChunkMetadata do
+  @moduledoc "Metadata about a chunk's position in the original document."
+
+  @typedoc "Metadata about a chunk's position in the original document."
+  @type t :: %__MODULE__{
+          byte_start: non_neg_integer(),
+          byte_end: non_neg_integer(),
+          token_count: non_neg_integer() | nil,
+          chunk_index: non_neg_integer(),
+          total_chunks: non_neg_integer(),
+          first_page: non_neg_integer() | nil,
+          last_page: non_neg_integer() | nil,
+          heading_context: map() | nil,
+          image_indices: [non_neg_integer()]
+        }
+
+  defstruct byte_start: 0,
+            byte_end: 0,
+            token_count: nil,
+            chunk_index: 0,
+            total_chunks: 0,
+            first_page: nil,
+            last_page: nil,
+            heading_context: nil,
+            image_indices: []
+end

packages/elixir/lib/kreuzberg/chunk_sizing.ex

@@ -0,0 +1,25 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ChunkSizing do
+  @moduledoc """
+  How chunk size is measured.
+
+  Defaults to `Characters` (Unicode character count). When using token-based sizing,
+  chunks are sized by token count according to the specified tokenizer.
+
+  Token-based sizing uses HuggingFace tokenizers loaded at runtime. Any tokenizer
+  available on HuggingFace Hub can be used, including OpenAI-compatible tokenizers
+  (e.g., `Xenova/gpt-4o`, `Xenova/cl100k_base`).
+  """
+
+  @typedoc "How chunk size is measured."
+  @type t :: term()
+
+  @typedoc "Size measured in Unicode characters (default)."
+  @type characters :: :characters
+  @typedoc "Size measured in tokens from a HuggingFace tokenizer."
+  @type tokenizer :: %{type: :tokenizer, model: String.t(), cache_dir: String.t()}
+end

packages/elixir/lib/kreuzberg/chunk_type.ex

@@ -0,0 +1,84 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ChunkType do
+  @moduledoc """
+  Semantic structural classification of a text chunk.
+
+  Assigned by the heuristic classifier in `chunking::classifier`.
+  Defaults to `Unknown` when no rule matches.
+  Designed to be extended in future versions without breaking changes.
+  """
+
+  @typedoc "Semantic structural classification of a text chunk."
+  @type t ::
+          :heading
+          | :party_list
+          | :definitions
+          | :operative_clause
+          | :signature_block
+          | :schedule
+          | :table_like
+          | :formula
+          | :code_block
+          | :image
+          | :org_chart
+          | :diagram
+          | :unknown
+
+  @heading :heading
+  @party_list :party_list
+  @definitions :definitions
+  @operative_clause :operative_clause
+  @signature_block :signature_block
+  @schedule :schedule
+  @table_like :table_like
+  @formula :formula
+  @code_block :code_block
+  @image :image
+  @org_chart :org_chart
+  @diagram :diagram
+  @unknown :unknown
+
+  @doc "Section heading or document title."
+  @spec heading() :: t()
+  def heading, do: @heading
+  @doc "Party list: names, addresses, and signatories."
+  @spec party_list() :: t()
+  def party_list, do: @party_list
+  @doc "Definition clause (\"X means…\", \"X shall mean…\")."
+  @spec definitions() :: t()
+  def definitions, do: @definitions
+  @doc "Operative clause containing legal/contractual action verbs."
+  @spec operative_clause() :: t()
+  def operative_clause, do: @operative_clause
+  @doc "Signature block with signatures, names, and dates."
+  @spec signature_block() :: t()
+  def signature_block, do: @signature_block
+  @doc "Schedule, annex, appendix, or exhibit section."
+  @spec schedule() :: t()
+  def schedule, do: @schedule
+  @doc "Table-like content with aligned columns or repeated patterns."
+  @spec table_like() :: t()
+  def table_like, do: @table_like
+  @doc "Mathematical formula or equation."
+  @spec formula() :: t()
+  def formula, do: @formula
+  @doc "Code block or preformatted content."
+  @spec code_block() :: t()
+  def code_block, do: @code_block
+  @doc "Embedded or referenced image content."
+  @spec image() :: t()
+  def image, do: @image
+  @doc "Organizational chart or hierarchy diagram."
+  @spec org_chart() :: t()
+  def org_chart, do: @org_chart
+  @doc "Diagram, figure, or visual illustration."
+  @spec diagram() :: t()
+  def diagram, do: @diagram
+  @doc "Unclassified or mixed content."
+  @spec unknown() :: t()
+  def unknown, do: @unknown
+end

packages/elixir/lib/kreuzberg/chunker_type.ex

@@ -0,0 +1,40 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ChunkerType do
+  @moduledoc """
+  Type of text chunker to use.
+
+  # Variants
+
+  * `Text` - Generic text splitter, splits on whitespace and punctuation
+  * `Markdown` - Markdown-aware splitter, preserves formatting and structure
+  * `Yaml` - YAML-aware splitter, creates one chunk per top-level key
+  * `Semantic` - Topic-aware chunker. With an `EmbeddingConfig`, splits at
+    embedding-based topic shifts tuned by `topic_threshold` (default 0.75,
+    lower = more splits). Without an embedding, falls back to a
+    structural-boundary heuristic (ALL-CAPS headers, numbered sections,
+    blank-line paragraphs) and merges groups into chunks capped at
+    `max_characters` (default 1000). `topic_threshold` has no effect in the
+    fallback path. For best results, pair with an embedding model.
+  """
+
+  @typedoc "Type of text chunker to use."
+  @type t :: :text | :markdown | :yaml | :semantic
+
+  @text :text
+  @markdown :markdown
+  @yaml :yaml
+  @semantic :semantic
+
+  @spec text() :: t()
+  def text, do: @text
+  @spec markdown() :: t()
+  def markdown, do: @markdown
+  @spec yaml() :: t()
+  def yaml, do: @yaml
+  @spec semantic() :: t()
+  def semantic, do: @semantic
+end

packages/elixir/lib/kreuzberg/chunking_config.ex

@@ -0,0 +1,55 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ChunkingConfig do
+  @moduledoc """
+  Chunking configuration.
+
+  Configures text chunking for document content, including chunk size,
+  overlap, trimming behavior, and optional embeddings.
+
+  Use `..Default::default()` when constructing to allow for future field additions:
+  ```rust
+  let config = ChunkingConfig {
+      max_characters: 500,
+      ..Default::default()
+  };
+  ```
+  """
+
+  @typedoc "Chunking configuration."
+  @type t :: %__MODULE__{
+          max_characters: non_neg_integer(),
+          overlap: non_neg_integer(),
+          trim: boolean(),
+          chunker_type: String.t() | nil,
+          embedding: map() | nil,
+          preset: String.t() | nil,
+          sizing: String.t() | nil,
+          prepend_heading_context: boolean(),
+          topic_threshold: float() | nil
+        }
+
+  defstruct max_characters: 1_000,
+            overlap: 200,
+            trim: true,
+            chunker_type: :text,
+            embedding: nil,
+            preset: nil,
+            sizing: :characters,
+            prepend_heading_context: false,
+            topic_threshold: nil
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/citation_metadata.ex

@@ -0,0 +1,36 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.CitationMetadata do
+  @moduledoc "Citation file metadata (RIS, PubMed, EndNote)."
+
+  @typedoc "Citation file metadata (RIS, PubMed, EndNote)."
+  @type t :: %__MODULE__{
+          citation_count: non_neg_integer(),
+          format: String.t() | nil,
+          authors: [String.t()],
+          year_range: map() | nil,
+          dois: [String.t()],
+          keywords: [String.t()]
+        }
+
+  defstruct citation_count: 0,
+            format: nil,
+            authors: [],
+            year_range: nil,
+            dois: [],
+            keywords: []
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/code_content_mode.ex

@@ -0,0 +1,30 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.CodeContentMode do
+  @moduledoc """
+  Content rendering mode for code extraction.
+
+  Controls how extracted code content is represented in the `content` field
+  of `ExtractionResult`.
+  """
+
+  @typedoc "Content rendering mode for code extraction."
+  @type t :: :chunks | :raw | :structure
+
+  @chunks :chunks
+  @raw :raw
+  @structure :structure
+
+  @doc "Use TSLP semantic chunks as content (default)."
+  @spec chunks() :: t()
+  def chunks, do: @chunks
+  @doc "Use raw source code as content."
+  @spec raw() :: t()
+  def raw, do: @raw
+  @doc "Emit function/class headings + docstrings (no code bodies)."
+  @spec structure() :: t()
+  def structure, do: @structure
+end

packages/elixir/lib/kreuzberg/content_filter_config.ex

@@ -0,0 +1,42 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ContentFilterConfig do
+  @moduledoc """
+  Cross-extractor content filtering configuration.
+
+  Controls whether "furniture" content (headers, footers, page numbers,
+  watermarks, repeating text) is included in or stripped from extraction
+  results. Applies across all extractors (PDF, DOCX, RTF, ODT, HTML, etc.)
+  with format-specific implementation.
+
+  When `None` on `ExtractionConfig`, each extractor uses its current
+  default behavior unchanged.
+  """
+
+  @typedoc "Cross-extractor content filtering configuration."
+  @type t :: %__MODULE__{
+          include_headers: boolean(),
+          include_footers: boolean(),
+          strip_repeating_text: boolean(),
+          include_watermarks: boolean()
+        }
+
+  defstruct include_headers: false,
+            include_footers: false,
+            strip_repeating_text: true,
+            include_watermarks: false
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/content_layer.ex

@@ -0,0 +1,33 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ContentLayer do
+  @moduledoc """
+  Content layer classification for document nodes.
+
+  Replaces separate body/furniture arrays with per-node granularity.
+  """
+
+  @typedoc "Content layer classification for document nodes."
+  @type t :: :body | :header | :footer | :footnote
+
+  @body :body
+  @header :header
+  @footer :footer
+  @footnote :footnote
+
+  @doc "Main document body content."
+  @spec body() :: t()
+  def body, do: @body
+  @doc "Page/section header (running header)."
+  @spec header() :: t()
+  def header, do: @header
+  @doc "Page/section footer (running footer)."
+  @spec footer() :: t()
+  def footer, do: @footer
+  @doc "Footnote content."
+  @spec footnote() :: t()
+  def footnote, do: @footnote
+end

packages/elixir/lib/kreuzberg/contributor_role.ex

@@ -0,0 +1,17 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ContributorRole do
+  @moduledoc "JATS contributor with role."
+
+  @typedoc "JATS contributor with role."
+  @type t :: %__MODULE__{
+          name: String.t() | nil,
+          role: String.t() | nil
+        }
+
+  defstruct name: nil,
+            role: nil
+end

packages/elixir/lib/kreuzberg/core_properties.ex

@@ -0,0 +1,59 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.CoreProperties do
+  @moduledoc """
+  Dublin Core metadata from docProps/core.xml
+
+  Contains standard metadata fields defined by the Dublin Core standard
+  and Office-specific extensions.
+  """
+
+  @typedoc "Dublin Core metadata from docProps/core.xml"
+  @type t :: %__MODULE__{
+          title: String.t() | nil,
+          subject: String.t() | nil,
+          creator: String.t() | nil,
+          keywords: String.t() | nil,
+          description: String.t() | nil,
+          last_modified_by: String.t() | nil,
+          revision: String.t() | nil,
+          created: String.t() | nil,
+          modified: String.t() | nil,
+          category: String.t() | nil,
+          content_status: String.t() | nil,
+          language: String.t() | nil,
+          identifier: String.t() | nil,
+          version: String.t() | nil,
+          last_printed: String.t() | nil
+        }
+
+  defstruct title: nil,
+            subject: nil,
+            creator: nil,
+            keywords: nil,
+            description: nil,
+            last_modified_by: nil,
+            revision: nil,
+            created: nil,
+            modified: nil,
+            category: nil,
+            content_status: nil,
+            language: nil,
+            identifier: nil,
+            version: nil,
+            last_printed: nil
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/csv_metadata.ex

@@ -0,0 +1,34 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.CsvMetadata do
+  @moduledoc "CSV/TSV file metadata."
+
+  @typedoc "CSV/TSV file metadata."
+  @type t :: %__MODULE__{
+          row_count: non_neg_integer(),
+          column_count: non_neg_integer(),
+          delimiter: String.t() | nil,
+          has_header: boolean(),
+          column_types: [String.t()] | nil
+        }
+
+  defstruct row_count: 0,
+            column_count: 0,
+            delimiter: nil,
+            has_header: false,
+            column_types: nil
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/dbf_field_info.ex

@@ -0,0 +1,17 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.DbfFieldInfo do
+  @moduledoc "dBASE field information."
+
+  @typedoc "dBASE field information."
+  @type t :: %__MODULE__{
+          name: String.t() | nil,
+          field_type: String.t() | nil
+        }
+
+  defstruct name: nil,
+            field_type: nil
+end

packages/elixir/lib/kreuzberg/dbf_metadata.ex

@@ -0,0 +1,30 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.DbfMetadata do
+  @moduledoc "dBASE (DBF) file metadata."
+
+  @typedoc "dBASE (DBF) file metadata."
+  @type t :: %__MODULE__{
+          record_count: non_neg_integer(),
+          field_count: non_neg_integer(),
+          fields: [map()]
+        }
+
+  defstruct record_count: 0,
+            field_count: 0,
+            fields: []
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/detect_response.ex

@@ -0,0 +1,17 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.DetectResponse do
+  @moduledoc "MIME type detection response."
+
+  @typedoc "MIME type detection response."
+  @type t :: %__MODULE__{
+          mime_type: String.t() | nil,
+          filename: String.t() | nil
+        }
+
+  defstruct mime_type: nil,
+            filename: nil
+end

packages/elixir/lib/kreuzberg/detection_result.ex

@@ -0,0 +1,19 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.DetectionResult do
+  @moduledoc "Page-level detection result containing all detections and page metadata."
+
+  @typedoc "Page-level detection result containing all detections and page metadata."
+  @type t :: %__MODULE__{
+          page_width: non_neg_integer(),
+          page_height: non_neg_integer(),
+          detections: [map()]
+        }
+
+  defstruct page_width: 0,
+            page_height: 0,
+            detections: []
+end

packages/elixir/lib/kreuzberg/diff_hunk.ex

@@ -0,0 +1,23 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.DiffHunk do
+  @moduledoc "A single contiguous hunk in a unified diff."
+
+  @typedoc "A single contiguous hunk in a unified diff."
+  @type t :: %__MODULE__{
+          from_line: non_neg_integer(),
+          from_count: non_neg_integer(),
+          to_line: non_neg_integer(),
+          to_count: non_neg_integer(),
+          lines: [String.t() | nil]
+        }
+
+  defstruct from_line: 0,
+            from_count: 0,
+            to_line: 0,
+            to_count: 0,
+            lines: []
+end

packages/elixir/lib/kreuzberg/diff_line.ex

@@ -0,0 +1,24 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.DiffLine do
+  @moduledoc """
+  A single line in a unified-diff hunk.
+
+  Defined here (rather than only in `crate::diff`) so `RevisionDelta` can
+  reference it unconditionally, without requiring the `diff` Cargo feature.
+  `crate::diff` re-exports this type verbatim.
+  """
+
+  @typedoc "A single line in a unified-diff hunk."
+  @type t :: term()
+
+  @typedoc "Unchanged context line."
+  @type context :: %{type: :context, value: String.t()}
+  @typedoc "Line added in the \"after\" version."
+  @type added :: %{type: :added, value: String.t()}
+  @typedoc "Line removed from the \"before\" version."
+  @type removed :: %{type: :removed, value: String.t()}
+end

packages/elixir/lib/kreuzberg/diff_options.ex

@@ -0,0 +1,30 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.DiffOptions do
+  @moduledoc "Options controlling how two `ExtractionResult` values are compared."
+
+  @typedoc "Options controlling how two `ExtractionResult` values are compared."
+  @type t :: %__MODULE__{
+          include_metadata: boolean(),
+          include_embedded: boolean(),
+          max_content_chars: non_neg_integer() | nil
+        }
+
+  defstruct include_metadata: true,
+            include_embedded: true,
+            max_content_chars: nil
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/djot_content.ex

@@ -0,0 +1,41 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.DjotContent do
+  @moduledoc """
+  Comprehensive Djot document structure with semantic preservation.
+
+  This type captures the full richness of Djot markup, including:
+  - Block-level structures (headings, lists, blockquotes, code blocks, etc.)
+  - Inline formatting (emphasis, strong, highlight, subscript, superscript, etc.)
+  - Attributes (classes, IDs, key-value pairs)
+  - Links, images, footnotes
+  - Math expressions (inline and display)
+  - Tables with full structure
+
+  Available when the `djot` feature is enabled.
+  """
+
+  @typedoc "Comprehensive Djot document structure with semantic preservation."
+  @type t :: %__MODULE__{
+          plain_text: String.t() | nil,
+          blocks: [map()],
+          metadata: map(),
+          tables: [map()],
+          images: [map()],
+          links: [map()],
+          footnotes: [map()],
+          attributes: [String.t()]
+        }
+
+  defstruct plain_text: nil,
+            blocks: [],
+            metadata: nil,
+            tables: [],
+            images: [],
+            links: [],
+            footnotes: [],
+            attributes: []
+end

packages/elixir/lib/kreuzberg/djot_image.ex

@@ -0,0 +1,21 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.DjotImage do
+  @moduledoc "Image element in Djot."
+
+  @typedoc "Image element in Djot."
+  @type t :: %__MODULE__{
+          src: String.t() | nil,
+          alt: String.t() | nil,
+          title: String.t() | nil,
+          attributes: String.t() | nil
+        }
+
+  defstruct src: nil,
+            alt: nil,
+            title: nil,
+            attributes: nil
+end

packages/elixir/lib/kreuzberg/djot_link.ex

@@ -0,0 +1,21 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.DjotLink do
+  @moduledoc "Link element in Djot."
+
+  @typedoc "Link element in Djot."
+  @type t :: %__MODULE__{
+          url: String.t() | nil,
+          text: String.t() | nil,
+          title: String.t() | nil,
+          attributes: String.t() | nil
+        }
+
+  defstruct url: nil,
+            text: nil,
+            title: nil,
+            attributes: nil
+end

packages/elixir/lib/kreuzberg/document_extractor_bridge.ex

@@ -0,0 +1,74 @@
+defmodule KreuzbergDocumentExtractorBridge do
+  @moduledoc """
+  GenServer bridge for DocumentExtractor implementation in kreuzberg.
+
+  Handles incoming trait method calls from Rust and dispatches them to an implementation module.
+  """
+
+  use GenServer
+
+  require Logger
+
+  @doc """
+  Start a GenServer linked to the current process.
+
+  impl_module should be a module that implements the DocumentExtractor trait methods.
+  """
+  def start_link(impl_module) do
+    GenServer.start_link(__MODULE__, impl_module, name: __MODULE__)
+  end
+
+  @impl GenServer
+  def init(impl_module) do
+    {:ok, impl_module}
+  end
+
+  @doc """
+  Handle an incoming trait call message.
+
+  Message format: {:trait_call, method_atom, args_json, reply_id}
+  """
+  @impl GenServer
+  def handle_info({:trait_call, method, args_json, reply_id}, impl_module) do
+    try do
+      args = Jason.decode!(args_json)
+      method_name = to_string(method)
+      ordered_args = ordered_args(impl_module, method_name, args)
+
+      # Dispatch to the implementation module
+      result = apply(impl_module, String.to_existing_atom(method_name), ordered_args)
+
+      # Send result back to Rust
+      Kreuzberg.Native.complete_trait_call(reply_id, Jason.encode!(result))
+    rescue
+      e ->
+        Logger.error("Error calling {impl_module}.{method}: {Exception.message(e)}")
+        Kreuzberg.Native.fail_trait_call(reply_id, Exception.message(e))
+    end
+
+    {:noreply, impl_module}
+  end
+
+  defp ordered_args(impl_module, method_name, args) when is_map(args) do
+    if function_exported?(impl_module, :__alef_arg_order__, 1) do
+      impl_module.__alef_arg_order__(method_name)
+      |> Enum.map(&Map.fetch!(args, &1))
+    else
+      args
+      |> Map.keys()
+      |> Enum.sort()
+      |> Enum.map(&Map.fetch!(args, &1))
+    end
+  end
+
+  defp ordered_args(_impl_module, _method_name, args) when is_list(args), do: args
+
+  @doc """
+  Register an implementation module, starting a GenServer to handle trait calls.
+  """
+  def register(impl_module) do
+    plugin_name = impl_module.name()
+    {:ok, pid} = start_link(impl_module)
+    Kreuzberg.Native.register_document_extractor(pid, plugin_name)
+  end
+end

packages/elixir/lib/kreuzberg/document_node.ex

@@ -0,0 +1,38 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.DocumentNode do
+  @moduledoc """
+  A single node in the document tree.
+
+  Each node has deterministic `id`, typed `content`, optional `parent`/`children`
+  for tree structure, and metadata like page number, bounding box, and content layer.
+  """
+
+  @typedoc "A single node in the document tree."
+  @type t :: %__MODULE__{
+          id: String.t() | nil,
+          content: String.t() | nil,
+          parent: non_neg_integer() | nil,
+          children: [non_neg_integer()],
+          content_layer: String.t() | nil,
+          page: non_neg_integer() | nil,
+          page_end: non_neg_integer() | nil,
+          bbox: map() | nil,
+          annotations: [map()],
+          attributes: map() | nil
+        }
+
+  defstruct id: nil,
+            content: :title,
+            parent: nil,
+            children: [],
+            content_layer: :body,
+            page: nil,
+            page_end: nil,
+            bbox: nil,
+            annotations: [],
+            attributes: nil
+end

packages/elixir/lib/kreuzberg/document_relationship.ex

@@ -0,0 +1,19 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.DocumentRelationship do
+  @moduledoc "A resolved relationship between two nodes in the document tree."
+
+  @typedoc "A resolved relationship between two nodes in the document tree."
+  @type t :: %__MODULE__{
+          source: non_neg_integer(),
+          target: non_neg_integer(),
+          kind: String.t() | nil
+        }
+
+  defstruct source: 0,
+            target: 0,
+            kind: :footnote_reference
+end

packages/elixir/lib/kreuzberg/document_revision.ex

@@ -0,0 +1,32 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.DocumentRevision do
+  @moduledoc """
+  A single tracked change embedded in a document.
+
+  Populated by per-format extractors that understand change-tracking metadata
+  (DOCX `w:ins`/`w:del`/`w:rPrChange`, ODT `text:change-*`, …). Every
+  extractor defaults to `ExtractionResult.revisions = None` until a
+  format-specific implementation is added.
+  """
+
+  @typedoc "A single tracked change embedded in a document."
+  @type t :: %__MODULE__{
+          revision_id: String.t() | nil,
+          author: String.t() | nil,
+          timestamp: String.t() | nil,
+          kind: String.t() | nil,
+          anchor: String.t() | nil | nil,
+          delta: map()
+        }
+
+  defstruct revision_id: nil,
+            author: nil,
+            timestamp: nil,
+            kind: :insertion,
+            anchor: nil,
+            delta: nil
+end

packages/elixir/lib/kreuzberg/document_structure.ex

@@ -0,0 +1,43 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.DocumentStructure do
+  @moduledoc """
+  Top-level structured document representation.
+
+  A flat array of nodes with index-based parent/child references forming a tree.
+  Root-level nodes have `parent: None`. Use `body_roots()` and `furniture_roots()`
+  to iterate over top-level content by layer.
+
+  # Validation
+
+  Call `validate()` after construction to verify all node indices are in bounds
+  and parent-child relationships are bidirectionally consistent.
+  """
+
+  @typedoc "Top-level structured document representation."
+  @type t :: %__MODULE__{
+          nodes: [map()],
+          source_format: String.t() | nil,
+          relationships: [map()],
+          node_types: [String.t()]
+        }
+
+  defstruct nodes: [],
+            source_format: nil,
+            relationships: [],
+            node_types: []
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/docx_app_properties.ex

@@ -0,0 +1,60 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.DocxAppProperties do
+  @moduledoc """
+  Application properties from docProps/app.xml for DOCX
+
+  Contains Word-specific document statistics and metadata.
+  """
+
+  @typedoc "Application properties from docProps/app.xml for DOCX"
+  @type t :: %__MODULE__{
+          application: String.t() | nil,
+          app_version: String.t() | nil,
+          template: String.t() | nil,
+          total_time: integer() | nil,
+          pages: integer() | nil,
+          words: integer() | nil,
+          characters: integer() | nil,
+          characters_with_spaces: integer() | nil,
+          lines: integer() | nil,
+          paragraphs: integer() | nil,
+          company: String.t() | nil,
+          doc_security: integer() | nil,
+          scale_crop: boolean() | nil,
+          links_up_to_date: boolean() | nil,
+          shared_doc: boolean() | nil,
+          hyperlinks_changed: boolean() | nil
+        }
+
+  defstruct application: nil,
+            app_version: nil,
+            template: nil,
+            total_time: nil,
+            pages: nil,
+            words: nil,
+            characters: nil,
+            characters_with_spaces: nil,
+            lines: nil,
+            paragraphs: nil,
+            company: nil,
+            doc_security: nil,
+            scale_crop: nil,
+            links_up_to_date: nil,
+            shared_doc: nil,
+            hyperlinks_changed: nil
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/docx_metadata.ex

@@ -0,0 +1,35 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.DocxMetadata do
+  @moduledoc """
+  Word document metadata.
+
+  Extracted from DOCX files using shared Office Open XML metadata extraction.
+  Integrates with `office_metadata` module for core/app/custom properties.
+  """
+
+  @typedoc "Word document metadata."
+  @type t :: %__MODULE__{
+          core_properties: map() | nil,
+          app_properties: map() | nil,
+          custom_properties: map() | nil
+        }
+
+  defstruct core_properties: nil,
+            app_properties: nil,
+            custom_properties: nil
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/element.ex

@@ -0,0 +1,26 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.Element do
+  @moduledoc """
+  Semantic element extracted from document.
+
+  Represents a logical unit of content with semantic classification,
+  unique identifier, and metadata for tracking origin and position.
+  """
+
+  @typedoc "Semantic element extracted from document."
+  @type t :: %__MODULE__{
+          element_id: String.t() | nil,
+          element_type: String.t() | nil,
+          text: String.t() | nil,
+          metadata: map()
+        }
+
+  defstruct element_id: nil,
+            element_type: :title,
+            text: nil,
+            metadata: nil
+end

packages/elixir/lib/kreuzberg/element_metadata.ex

@@ -0,0 +1,23 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ElementMetadata do
+  @moduledoc "Metadata for a semantic element."
+
+  @typedoc "Metadata for a semantic element."
+  @type t :: %__MODULE__{
+          page_number: non_neg_integer() | nil,
+          filename: String.t() | nil,
+          coordinates: map() | nil,
+          element_index: non_neg_integer() | nil,
+          additional: map()
+        }
+
+  defstruct page_number: nil,
+            filename: nil,
+            coordinates: nil,
+            element_index: nil,
+            additional: %{}
+end

packages/elixir/lib/kreuzberg/element_type.ex

@@ -0,0 +1,73 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ElementType do
+  @moduledoc """
+  Semantic element type classification.
+
+  Categorizes text content into semantic units for downstream processing.
+  Supports the element types commonly found in Unstructured documents.
+  """
+
+  @typedoc "Semantic element type classification."
+  @type t ::
+          :title
+          | :narrative_text
+          | :heading
+          | :list_item
+          | :table
+          | :image
+          | :page_break
+          | :code_block
+          | :block_quote
+          | :footer
+          | :header
+
+  @title :title
+  @narrative_text :narrative_text
+  @heading :heading
+  @list_item :list_item
+  @table :table
+  @image :image
+  @page_break :page_break
+  @code_block :code_block
+  @block_quote :block_quote
+  @footer :footer
+  @header :header
+
+  @doc "Document title"
+  @spec title() :: t()
+  def title, do: @title
+  @doc "Main narrative text body"
+  @spec narrative_text() :: t()
+  def narrative_text, do: @narrative_text
+  @doc "Section heading"
+  @spec heading() :: t()
+  def heading, do: @heading
+  @doc "List item (bullet, numbered, etc.)"
+  @spec list_item() :: t()
+  def list_item, do: @list_item
+  @doc "Table element"
+  @spec table() :: t()
+  def table, do: @table
+  @doc "Image element"
+  @spec image() :: t()
+  def image, do: @image
+  @doc "Page break marker"
+  @spec page_break() :: t()
+  def page_break, do: @page_break
+  @doc "Code block"
+  @spec code_block() :: t()
+  def code_block, do: @code_block
+  @doc "Block quote"
+  @spec block_quote() :: t()
+  def block_quote, do: @block_quote
+  @doc "Footer text"
+  @spec footer() :: t()
+  def footer, do: @footer
+  @doc "Header text"
+  @spec header() :: t()
+  def header, do: @header
+end

packages/elixir/lib/kreuzberg/email_attachment.ex

@@ -0,0 +1,29 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.EmailAttachment do
+  @moduledoc """
+  Email attachment representation.
+
+  Contains metadata and optionally the content of an email attachment.
+  """
+
+  @typedoc "Email attachment representation."
+  @type t :: %__MODULE__{
+          name: String.t() | nil,
+          filename: String.t() | nil,
+          mime_type: String.t() | nil,
+          size: non_neg_integer() | nil,
+          is_image: boolean(),
+          data: binary() | nil
+        }
+
+  defstruct name: nil,
+            filename: nil,
+            mime_type: nil,
+            size: nil,
+            is_image: false,
+            data: nil
+end

packages/elixir/lib/kreuzberg/email_config.ex

@@ -0,0 +1,26 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.EmailConfig do
+  @moduledoc "Configuration for email extraction."
+
+  @typedoc "Configuration for email extraction."
+  @type t :: %__MODULE__{
+          msg_fallback_codepage: non_neg_integer() | nil
+        }
+
+  defstruct msg_fallback_codepage: nil
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/email_extraction_result.ex

@@ -0,0 +1,42 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.EmailExtractionResult do
+  @moduledoc """
+  Email extraction result.
+
+  Complete representation of an extracted email message (.eml or .msg)
+  including headers, body content, and attachments.
+  """
+
+  @typedoc "Email extraction result."
+  @type t :: %__MODULE__{
+          subject: String.t() | nil,
+          from_email: String.t() | nil,
+          to_emails: [String.t()],
+          cc_emails: [String.t()],
+          bcc_emails: [String.t()],
+          date: String.t() | nil,
+          message_id: String.t() | nil,
+          plain_text: String.t() | nil,
+          html_content: String.t() | nil,
+          content: String.t() | nil,
+          attachments: [map()],
+          metadata: map()
+        }
+
+  defstruct subject: nil,
+            from_email: nil,
+            to_emails: [],
+            cc_emails: [],
+            bcc_emails: [],
+            date: nil,
+            message_id: nil,
+            plain_text: nil,
+            html_content: nil,
+            content: nil,
+            attachments: [],
+            metadata: %{}
+end

packages/elixir/lib/kreuzberg/email_metadata.ex

@@ -0,0 +1,42 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.EmailMetadata do
+  @moduledoc """
+  Email metadata extracted from .eml and .msg files.
+
+  Includes sender/recipient information, message ID, and attachment list.
+  """
+
+  @typedoc "Email metadata extracted from .eml and .msg files."
+  @type t :: %__MODULE__{
+          from_email: String.t() | nil,
+          from_name: String.t() | nil,
+          to_emails: [String.t()],
+          cc_emails: [String.t()],
+          bcc_emails: [String.t()],
+          message_id: String.t() | nil,
+          attachments: [String.t()]
+        }
+
+  defstruct from_email: nil,
+            from_name: nil,
+            to_emails: [],
+            cc_emails: [],
+            bcc_emails: [],
+            message_id: nil,
+            attachments: []
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/embedded_changes.ex

@@ -0,0 +1,19 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.EmbeddedChanges do
+  @moduledoc "Changes to embedded archive children between two results."
+
+  @typedoc "Changes to embedded archive children between two results."
+  @type t :: %__MODULE__{
+          added: [map()],
+          removed: [map()],
+          changed: [map()]
+        }
+
+  defstruct added: [],
+            removed: [],
+            changed: []
+end

packages/elixir/lib/kreuzberg/embedded_diff.ex

@@ -0,0 +1,17 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.EmbeddedDiff do
+  @moduledoc "Diff for a single embedded archive entry that appears in both results."
+
+  @typedoc "Diff for a single embedded archive entry that appears in both results."
+  @type t :: %__MODULE__{
+          path: String.t() | nil,
+          diff: map()
+        }
+
+  defstruct path: nil,
+            diff: nil
+end

packages/elixir/lib/kreuzberg/embedded_file.ex

@@ -0,0 +1,21 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.EmbeddedFile do
+  @moduledoc "Embedded file descriptor extracted from the PDF name tree."
+
+  @typedoc "Embedded file descriptor extracted from the PDF name tree."
+  @type t :: %__MODULE__{
+          name: String.t() | nil,
+          data: binary(),
+          compressed_size: non_neg_integer(),
+          mime_type: String.t() | nil
+        }
+
+  defstruct name: nil,
+            data: <<>>,
+            compressed_size: 0,
+            mime_type: nil
+end

packages/elixir/lib/kreuzberg/embedding_backend_bridge.ex

@@ -0,0 +1,74 @@
+defmodule KreuzbergEmbeddingBackendBridge do
+  @moduledoc """
+  GenServer bridge for EmbeddingBackend implementation in kreuzberg.
+
+  Handles incoming trait method calls from Rust and dispatches them to an implementation module.
+  """
+
+  use GenServer
+
+  require Logger
+
+  @doc """
+  Start a GenServer linked to the current process.
+
+  impl_module should be a module that implements the EmbeddingBackend trait methods.
+  """
+  def start_link(impl_module) do
+    GenServer.start_link(__MODULE__, impl_module, name: __MODULE__)
+  end
+
+  @impl GenServer
+  def init(impl_module) do
+    {:ok, impl_module}
+  end
+
+  @doc """
+  Handle an incoming trait call message.
+
+  Message format: {:trait_call, method_atom, args_json, reply_id}
+  """
+  @impl GenServer
+  def handle_info({:trait_call, method, args_json, reply_id}, impl_module) do
+    try do
+      args = Jason.decode!(args_json)
+      method_name = to_string(method)
+      ordered_args = ordered_args(impl_module, method_name, args)
+
+      # Dispatch to the implementation module
+      result = apply(impl_module, String.to_existing_atom(method_name), ordered_args)
+
+      # Send result back to Rust
+      Kreuzberg.Native.complete_trait_call(reply_id, Jason.encode!(result))
+    rescue
+      e ->
+        Logger.error("Error calling {impl_module}.{method}: {Exception.message(e)}")
+        Kreuzberg.Native.fail_trait_call(reply_id, Exception.message(e))
+    end
+
+    {:noreply, impl_module}
+  end
+
+  defp ordered_args(impl_module, method_name, args) when is_map(args) do
+    if function_exported?(impl_module, :__alef_arg_order__, 1) do
+      impl_module.__alef_arg_order__(method_name)
+      |> Enum.map(&Map.fetch!(args, &1))
+    else
+      args
+      |> Map.keys()
+      |> Enum.sort()
+      |> Enum.map(&Map.fetch!(args, &1))
+    end
+  end
+
+  defp ordered_args(_impl_module, _method_name, args) when is_list(args), do: args
+
+  @doc """
+  Register an implementation module, starting a GenServer to handle trait calls.
+  """
+  def register(impl_module) do
+    plugin_name = impl_module.name()
+    {:ok, pid} = start_link(impl_module)
+    Kreuzberg.Native.register_embedding_backend(pid, plugin_name)
+  end
+end

packages/elixir/lib/kreuzberg/embedding_config.ex

@@ -0,0 +1,43 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.EmbeddingConfig do
+  @moduledoc """
+  Embedding configuration for text chunks.
+
+  Configures embedding generation using ONNX models via the vendored embedding engine.
+  Requires the `embeddings` feature to be enabled.
+  """
+
+  @typedoc "Embedding configuration for text chunks."
+  @type t :: %__MODULE__{
+          model: String.t() | nil,
+          normalize: boolean(),
+          batch_size: non_neg_integer(),
+          show_download_progress: boolean(),
+          cache_dir: String.t() | nil,
+          acceleration: map() | nil,
+          max_embed_duration_secs: non_neg_integer() | nil
+        }
+
+  defstruct model: :preset,
+            normalize: true,
+            batch_size: 32,
+            show_download_progress: false,
+            cache_dir: nil,
+            acceleration: nil,
+            max_embed_duration_secs: nil
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/embedding_model_type.ex

@@ -0,0 +1,20 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.EmbeddingModelType do
+  @moduledoc "Embedding model types supported by Kreuzberg."
+
+  @typedoc "Embedding model types supported by Kreuzberg."
+  @type t :: term()
+
+  @typedoc "Use a preset model configuration (recommended)"
+  @type preset :: %{type: :preset, name: String.t()}
+  @typedoc "Use a custom ONNX model from HuggingFace"
+  @type custom :: %{type: :custom, model_id: String.t(), dimensions: non_neg_integer()}
+  @typedoc "Provider-hosted embedding model via liter-llm."
+  @type llm :: %{type: :llm, llm: Kreuzberg.LlmConfig.t()}
+  @typedoc "In-process embedding backend registered via the plugin system."
+  @type plugin :: %{type: :plugin, name: String.t()}
+end

packages/elixir/lib/kreuzberg/embedding_preset.ex

@@ -0,0 +1,37 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.EmbeddingPreset do
+  @moduledoc """
+  Preset configurations for common RAG use cases.
+
+  Each preset combines chunk size, overlap, and embedding model
+  to provide an optimized configuration for specific scenarios.
+
+  All string fields are owned `String` for FFI compatibility — instances
+  are safe to clone and pass across language boundaries.
+  """
+
+  @typedoc "Preset configurations for common RAG use cases."
+  @type t :: %__MODULE__{
+          name: String.t() | nil,
+          chunk_size: non_neg_integer(),
+          overlap: non_neg_integer(),
+          model_repo: String.t() | nil,
+          pooling: String.t() | nil,
+          model_file: String.t() | nil,
+          dimensions: non_neg_integer(),
+          description: String.t() | nil
+        }
+
+  defstruct name: nil,
+            chunk_size: 0,
+            overlap: 0,
+            model_repo: nil,
+            pooling: nil,
+            model_file: nil,
+            dimensions: 0,
+            description: nil
+end

packages/elixir/lib/kreuzberg/epub_metadata.ex

@@ -0,0 +1,36 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.EpubMetadata do
+  @moduledoc "EPUB metadata (Dublin Core extensions)."
+
+  @typedoc "EPUB metadata (Dublin Core extensions)."
+  @type t :: %__MODULE__{
+          coverage: String.t() | nil,
+          dc_format: String.t() | nil,
+          relation: String.t() | nil,
+          source: String.t() | nil,
+          dc_type: String.t() | nil,
+          cover_image: String.t() | nil
+        }
+
+  defstruct coverage: nil,
+            dc_format: nil,
+            relation: nil,
+            source: nil,
+            dc_type: nil,
+            cover_image: nil
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/error_metadata.ex

@@ -0,0 +1,17 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ErrorMetadata do
+  @moduledoc "Error metadata (for batch operations)."
+
+  @typedoc "Error metadata (for batch operations)."
+  @type t :: %__MODULE__{
+          error_type: String.t() | nil,
+          message: String.t() | nil
+        }
+
+  defstruct error_type: nil,
+            message: nil
+end

packages/elixir/lib/kreuzberg/excel_metadata.ex

@@ -0,0 +1,33 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ExcelMetadata do
+  @moduledoc """
+  Excel/spreadsheet format metadata.
+
+  Identifies the document as a spreadsheet source via the `FormatMetadata::Excel`
+  discriminant. Sheet count and sheet names are stored inside this struct.
+  """
+
+  @typedoc "Excel/spreadsheet format metadata."
+  @type t :: %__MODULE__{
+          sheet_count: non_neg_integer() | nil,
+          sheet_names: [String.t()] | nil
+        }
+
+  defstruct sheet_count: nil,
+            sheet_names: nil
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/excel_sheet.ex

@@ -0,0 +1,30 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ExcelSheet do
+  @moduledoc """
+  Single Excel worksheet.
+
+  Represents one sheet from an Excel workbook with its content
+  converted to Markdown format and dimensional statistics.
+  """
+
+  @typedoc "Single Excel worksheet."
+  @type t :: %__MODULE__{
+          name: String.t() | nil,
+          markdown: String.t() | nil,
+          row_count: non_neg_integer(),
+          col_count: non_neg_integer(),
+          cell_count: non_neg_integer(),
+          table_cells: [[String.t()]] | nil
+        }
+
+  defstruct name: nil,
+            markdown: nil,
+            row_count: 0,
+            col_count: 0,
+            cell_count: 0,
+            table_cells: nil
+end

packages/elixir/lib/kreuzberg/excel_workbook.ex

@@ -0,0 +1,24 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ExcelWorkbook do
+  @moduledoc """
+  Excel workbook representation.
+
+  Contains all sheets from an Excel file (.xlsx, .xls, etc.) with
+  extracted content and metadata.
+  """
+
+  @typedoc "Excel workbook representation."
+  @type t :: %__MODULE__{
+          sheets: [map()],
+          metadata: map(),
+          revisions: [map()] | nil
+        }
+
+  defstruct sheets: [],
+            metadata: %{},
+            revisions: nil
+end

packages/elixir/lib/kreuzberg/execution_provider_type.ex

@@ -0,0 +1,38 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ExecutionProviderType do
+  @moduledoc """
+  ONNX Runtime execution provider type.
+
+  Determines which hardware backend is used for model inference.
+  `Auto` (default) selects the best available provider per platform.
+  """
+
+  @typedoc "ONNX Runtime execution provider type."
+  @type t :: :auto | :cpu | :core_ml | :cuda | :tensor_rt
+
+  @auto :auto
+  @cpu :cpu
+  @core_ml :core_ml
+  @cuda :cuda
+  @tensor_rt :tensor_rt
+
+  @doc "Auto-select: CoreML on macOS, CUDA on Linux, CPU elsewhere."
+  @spec auto() :: t()
+  def auto, do: @auto
+  @doc "CPU execution provider (always available)."
+  @spec cpu() :: t()
+  def cpu, do: @cpu
+  @doc "Apple CoreML (macOS/iOS Neural Engine + GPU)."
+  @spec core_ml() :: t()
+  def core_ml, do: @core_ml
+  @doc "NVIDIA CUDA GPU acceleration."
+  @spec cuda() :: t()
+  def cuda, do: @cuda
+  @doc "NVIDIA TensorRT (optimized CUDA inference)."
+  @spec tensor_rt() :: t()
+  def tensor_rt, do: @tensor_rt
+end

packages/elixir/lib/kreuzberg/extracted_image.ex

@@ -0,0 +1,51 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ExtractedImage do
+  @moduledoc """
+  Extracted image from a document.
+
+  Contains raw image data, metadata, and optional nested OCR results.
+  Raw bytes allow cross-language compatibility - users can convert to
+  PIL.Image (Python), Sharp (Node.js), or other formats as needed.
+  """
+
+  @typedoc "Extracted image from a document."
+  @type t :: %__MODULE__{
+          data: binary(),
+          format: String.t() | nil,
+          image_index: non_neg_integer(),
+          page_number: non_neg_integer() | nil,
+          width: non_neg_integer() | nil,
+          height: non_neg_integer() | nil,
+          colorspace: String.t() | nil,
+          bits_per_component: non_neg_integer() | nil,
+          is_mask: boolean(),
+          description: String.t() | nil,
+          ocr_result: map() | nil,
+          bounding_box: map() | nil,
+          source_path: String.t() | nil,
+          image_kind: String.t() | nil | nil,
+          kind_confidence: float() | nil,
+          cluster_id: non_neg_integer() | nil
+        }
+
+  defstruct data: <<>>,
+            format: nil,
+            image_index: 0,
+            page_number: nil,
+            width: nil,
+            height: nil,
+            colorspace: nil,
+            bits_per_component: nil,
+            is_mask: false,
+            description: nil,
+            ocr_result: nil,
+            bounding_box: nil,
+            source_path: nil,
+            image_kind: nil,
+            kind_confidence: nil,
+            cluster_id: nil
+end

packages/elixir/lib/kreuzberg/extracted_uri.ex

@@ -0,0 +1,27 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ExtractedUri do
+  @moduledoc """
+  A URI extracted from a document.
+
+  Represents any link, reference, or resource pointer found during extraction.
+  The `kind` field classifies the URI semantically, while `label` carries
+  optional human-readable display text.
+  """
+
+  @typedoc "A URI extracted from a document."
+  @type t :: %__MODULE__{
+          url: String.t() | nil,
+          label: String.t() | nil,
+          page: non_neg_integer() | nil,
+          kind: String.t() | nil
+        }
+
+  defstruct url: nil,
+            label: nil,
+            page: nil,
+            kind: :hyperlink
+end

packages/elixir/lib/kreuzberg/extraction_config.ex

@@ -0,0 +1,111 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ExtractionConfig do
+  @moduledoc """
+  Main extraction configuration.
+
+  This struct contains all configuration options for the extraction process.
+  It can be loaded from TOML, YAML, or JSON files, or created programmatically.
+
+  # Example
+
+  ```rust
+  use kreuzberg::core::config::ExtractionConfig;
+
+  // Create with defaults
+  let config = ExtractionConfig::default();
+
+  // Load from TOML file
+  // let config = ExtractionConfig::from_toml_file("kreuzberg.toml")?;
+  ```
+  """
+
+  @typedoc "Main extraction configuration."
+  @type t :: %__MODULE__{
+          use_cache: boolean(),
+          enable_quality_processing: boolean(),
+          ocr: map() | nil,
+          force_ocr: boolean(),
+          force_ocr_pages: [non_neg_integer()] | nil,
+          disable_ocr: boolean(),
+          chunking: map() | nil,
+          content_filter: map() | nil,
+          images: map() | nil,
+          pdf_options: map() | nil,
+          token_reduction: map() | nil,
+          language_detection: map() | nil,
+          pages: map() | nil,
+          keywords: map() | nil,
+          postprocessor: map() | nil,
+          html_options: String.t() | nil,
+          html_output: map() | nil,
+          extraction_timeout_secs: non_neg_integer() | nil,
+          max_concurrent_extractions: non_neg_integer() | nil,
+          result_format: String.t() | nil,
+          security_limits: map() | nil,
+          max_embedded_file_bytes: non_neg_integer() | nil,
+          output_format: String.t() | nil,
+          layout: map() | nil,
+          use_layout_for_markdown: boolean(),
+          include_document_structure: boolean(),
+          acceleration: map() | nil,
+          cache_namespace: String.t() | nil,
+          cache_ttl_secs: non_neg_integer() | nil,
+          email: map() | nil,
+          concurrency: String.t() | nil,
+          max_archive_depth: non_neg_integer(),
+          tree_sitter: map() | nil,
+          structured_extraction: map() | nil,
+          cancel_token: String.t() | nil
+        }
+
+  defstruct use_cache: true,
+            enable_quality_processing: true,
+            ocr: nil,
+            force_ocr: false,
+            force_ocr_pages: nil,
+            disable_ocr: false,
+            chunking: nil,
+            content_filter: nil,
+            images: nil,
+            pdf_options: nil,
+            token_reduction: nil,
+            language_detection: nil,
+            pages: nil,
+            keywords: nil,
+            postprocessor: nil,
+            html_options: nil,
+            html_output: nil,
+            extraction_timeout_secs: nil,
+            max_concurrent_extractions: nil,
+            result_format: :unified,
+            security_limits: nil,
+            max_embedded_file_bytes: nil,
+            output_format: :plain,
+            layout: nil,
+            use_layout_for_markdown: false,
+            include_document_structure: false,
+            acceleration: nil,
+            cache_namespace: nil,
+            cache_ttl_secs: nil,
+            email: nil,
+            concurrency: nil,
+            max_archive_depth: 0,
+            tree_sitter: nil,
+            structured_extraction: nil,
+            cancel_token: nil
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/extraction_diff.ex

@@ -0,0 +1,25 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ExtractionDiff do
+  @moduledoc "The complete diff between two `ExtractionResult` values."
+
+  @typedoc "The complete diff between two `ExtractionResult` values."
+  @type t :: %__MODULE__{
+          content_diff: [map()],
+          tables_added: [map()],
+          tables_removed: [map()],
+          tables_changed: [map()],
+          metadata_changed: String.t() | nil,
+          embedded_changes: map()
+        }
+
+  defstruct content_diff: [],
+            tables_added: [],
+            tables_removed: [],
+            tables_changed: [],
+            metadata_changed: nil,
+            embedded_changes: nil
+end

packages/elixir/lib/kreuzberg/extraction_method.ex

@@ -0,0 +1,22 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ExtractionMethod do
+  @moduledoc "How the extracted text was produced."
+
+  @typedoc "How the extracted text was produced."
+  @type t :: :native | :ocr | :mixed
+
+  @native :native
+  @ocr :ocr
+  @mixed :mixed
+
+  @spec native() :: t()
+  def native, do: @native
+  @spec ocr() :: t()
+  def ocr, do: @ocr
+  @spec mixed() :: t()
+  def mixed, do: @mixed
+end

packages/elixir/lib/kreuzberg/extraction_result.ex

@@ -0,0 +1,78 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ExtractionResult do
+  @moduledoc """
+  General extraction result used by the core extraction API.
+
+  This is the main result type returned by all extraction functions.
+  """
+
+  @typedoc "General extraction result used by the core extraction API."
+  @type t :: %__MODULE__{
+          content: String.t() | nil,
+          mime_type: String.t() | nil,
+          metadata: map(),
+          extraction_method: String.t() | nil | nil,
+          tables: [map()],
+          detected_languages: [String.t()] | nil,
+          chunks: [map()] | nil,
+          images: [map()] | nil,
+          pages: [map()] | nil,
+          elements: [map()] | nil,
+          djot_content: map() | nil,
+          ocr_elements: [map()] | nil,
+          document: map() | nil,
+          extracted_keywords: [map()] | nil,
+          quality_score: float() | nil,
+          processing_warnings: [map()],
+          annotations: [map()] | nil,
+          children: [map()] | nil,
+          uris: [map()] | nil,
+          revisions: [map()] | nil,
+          structured_output: String.t() | nil,
+          code_intelligence: String.t() | nil,
+          llm_usage: [map()] | nil,
+          formatted_content: String.t() | nil,
+          ocr_internal_document: String.t() | nil
+        }
+
+  defstruct content: nil,
+            mime_type: nil,
+            metadata: nil,
+            extraction_method: nil,
+            tables: [],
+            detected_languages: nil,
+            chunks: nil,
+            images: nil,
+            pages: nil,
+            elements: nil,
+            djot_content: nil,
+            ocr_elements: nil,
+            document: nil,
+            extracted_keywords: nil,
+            quality_score: nil,
+            processing_warnings: [],
+            annotations: nil,
+            children: nil,
+            uris: nil,
+            revisions: nil,
+            structured_output: nil,
+            code_intelligence: nil,
+            llm_usage: nil,
+            formatted_content: nil,
+            ocr_internal_document: nil
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/fiction_book_metadata.ex

@@ -0,0 +1,30 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.FictionBookMetadata do
+  @moduledoc "FictionBook (FB2) metadata."
+
+  @typedoc "FictionBook (FB2) metadata."
+  @type t :: %__MODULE__{
+          genres: [String.t()],
+          sequences: [String.t()],
+          annotation: String.t() | nil
+        }
+
+  defstruct genres: [],
+            sequences: [],
+            annotation: nil
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/file_extraction_config.ex

@@ -0,0 +1,96 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.FileExtractionConfig do
+  @moduledoc """
+  Per-file extraction configuration overrides for batch processing.
+
+  All fields are `Option<T>` — `None` means "use the batch-level default."
+  This type is used with `batch_extract_files` and
+  `batch_extract_bytes` to allow heterogeneous
+  extraction settings within a single batch.
+
+  # Excluded Fields
+
+  The following `ExtractionConfig` fields are batch-level only and
+  cannot be overridden per file:
+  - `max_concurrent_extractions` — controls batch parallelism
+  - `use_cache` — global caching policy
+  - `acceleration` — shared ONNX execution provider
+  - `security_limits` — global archive security policy
+
+  # Example
+
+  ```rust
+  use kreuzberg::FileExtractionConfig;
+
+  // Override just OCR forcing for a specific file
+  let config = FileExtractionConfig {
+      force_ocr: Some(true),
+      ..Default::default()
+  };
+  ```
+  """
+
+  @typedoc "Per-file extraction configuration overrides for batch processing."
+  @type t :: %__MODULE__{
+          enable_quality_processing: boolean() | nil,
+          ocr: map() | nil,
+          force_ocr: boolean() | nil,
+          force_ocr_pages: [non_neg_integer()] | nil,
+          disable_ocr: boolean() | nil,
+          chunking: map() | nil,
+          content_filter: map() | nil,
+          images: map() | nil,
+          pdf_options: map() | nil,
+          token_reduction: map() | nil,
+          language_detection: map() | nil,
+          pages: map() | nil,
+          keywords: map() | nil,
+          postprocessor: map() | nil,
+          html_options: String.t() | nil,
+          result_format: String.t() | nil | nil,
+          output_format: String.t() | nil | nil,
+          include_document_structure: boolean() | nil,
+          layout: map() | nil,
+          timeout_secs: non_neg_integer() | nil,
+          tree_sitter: map() | nil,
+          structured_extraction: map() | nil
+        }
+
+  defstruct enable_quality_processing: nil,
+            ocr: nil,
+            force_ocr: nil,
+            force_ocr_pages: nil,
+            disable_ocr: nil,
+            chunking: nil,
+            content_filter: nil,
+            images: nil,
+            pdf_options: nil,
+            token_reduction: nil,
+            language_detection: nil,
+            pages: nil,
+            keywords: nil,
+            postprocessor: nil,
+            html_options: nil,
+            result_format: nil,
+            output_format: nil,
+            include_document_structure: nil,
+            layout: nil,
+            timeout_secs: nil,
+            tree_sitter: nil,
+            structured_extraction: nil
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/footnote.ex

@@ -0,0 +1,17 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.Footnote do
+  @moduledoc "Footnote in Djot."
+
+  @typedoc "Footnote in Djot."
+  @type t :: %__MODULE__{
+          label: String.t() | nil,
+          content: [map()]
+        }
+
+  defstruct label: nil,
+            content: []
+end

packages/elixir/lib/kreuzberg/format_metadata.ex

@@ -0,0 +1,37 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.FormatMetadata do
+  @moduledoc """
+  Format-specific metadata (discriminated union).
+
+  Only one format type can exist per extraction result. This provides
+  type-safe, clean metadata without nested optionals.
+  """
+
+  @typedoc "Format-specific metadata (discriminated union)."
+  @type t :: term()
+
+  @type pdf :: %{type: :pdf, metadata: Kreuzberg.PdfMetadata.t()}
+  @type docx :: %{type: :docx, metadata: Kreuzberg.DocxMetadata.t()}
+  @type excel :: %{type: :excel, metadata: Kreuzberg.ExcelMetadata.t()}
+  @type email :: %{type: :email, metadata: Kreuzberg.EmailMetadata.t()}
+  @type pptx :: %{type: :pptx, metadata: Kreuzberg.PptxMetadata.t()}
+  @type archive :: %{type: :archive, metadata: Kreuzberg.ArchiveMetadata.t()}
+  @type image :: %{type: :image, metadata: Kreuzberg.ImageMetadata.t()}
+  @type xml :: %{type: :xml, metadata: Kreuzberg.XmlMetadata.t()}
+  @type text :: %{type: :text, metadata: Kreuzberg.TextMetadata.t()}
+  @type html :: %{type: :html, metadata: Kreuzberg.HtmlMetadata.t()}
+  @type ocr :: %{type: :ocr, metadata: Kreuzberg.OcrMetadata.t()}
+  @type csv :: %{type: :csv, metadata: Kreuzberg.CsvMetadata.t()}
+  @type bibtex :: %{type: :bibtex, metadata: Kreuzberg.BibtexMetadata.t()}
+  @type citation :: %{type: :citation, metadata: Kreuzberg.CitationMetadata.t()}
+  @type fiction_book :: %{type: :fiction_book, metadata: Kreuzberg.FictionBookMetadata.t()}
+  @type dbf :: %{type: :dbf, metadata: Kreuzberg.DbfMetadata.t()}
+  @type jats :: %{type: :jats, metadata: Kreuzberg.JatsMetadata.t()}
+  @type epub :: %{type: :epub, metadata: Kreuzberg.EpubMetadata.t()}
+  @type pst :: %{type: :pst, metadata: Kreuzberg.PstMetadata.t()}
+  @type code :: %{type: :code, value: String.t()}
+end

packages/elixir/lib/kreuzberg/formatted_block.ex

@@ -0,0 +1,31 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.FormattedBlock do
+  @moduledoc """
+  Block-level element in a Djot document.
+
+  Represents structural elements like headings, paragraphs, lists, code blocks, etc.
+  """
+
+  @typedoc "Block-level element in a Djot document."
+  @type t :: %__MODULE__{
+          block_type: String.t() | nil,
+          level: non_neg_integer() | nil,
+          inline_content: [map()],
+          attributes: String.t() | nil,
+          language: String.t() | nil,
+          code: String.t() | nil,
+          children: [map()]
+        }
+
+  defstruct block_type: :paragraph,
+            level: nil,
+            inline_content: [],
+            attributes: nil,
+            language: nil,
+            code: nil,
+            children: []
+end

packages/elixir/lib/kreuzberg/grid_cell.ex

@@ -0,0 +1,27 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.GridCell do
+  @moduledoc "Individual grid cell with position and span metadata."
+
+  @typedoc "Individual grid cell with position and span metadata."
+  @type t :: %__MODULE__{
+          content: String.t() | nil,
+          row: non_neg_integer(),
+          col: non_neg_integer(),
+          row_span: non_neg_integer(),
+          col_span: non_neg_integer(),
+          is_header: boolean(),
+          bbox: map() | nil
+        }
+
+  defstruct content: nil,
+            row: 0,
+            col: 0,
+            row_span: 0,
+            col_span: 0,
+            is_header: false,
+            bbox: nil
+end

packages/elixir/lib/kreuzberg/header_metadata.ex

@@ -0,0 +1,29 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.HeaderMetadata do
+  @moduledoc "Header/heading element metadata."
+
+  @typedoc "Header/heading element metadata."
+  @type t :: %__MODULE__{
+          level: non_neg_integer(),
+          text: String.t() | nil,
+          id: String.t() | nil,
+          depth: non_neg_integer(),
+          html_offset: non_neg_integer()
+        }
+
+  defstruct level: 0,
+            text: nil,
+            id: nil,
+            depth: 0,
+            html_offset: 0
+
+  @doc "Validate that the header level is within valid range (1-6)."
+  @spec valid?(t()) :: boolean()
+  def valid?(%__MODULE__{level: level}) do
+    level >= 1 and level <= 6
+  end
+end

packages/elixir/lib/kreuzberg/heading_context.ex

@@ -0,0 +1,19 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.HeadingContext do
+  @moduledoc """
+  Heading context for a chunk within a Markdown document.
+
+  Contains the heading hierarchy from document root to this chunk's section.
+  """
+
+  @typedoc "Heading context for a chunk within a Markdown document."
+  @type t :: %__MODULE__{
+          headings: [map()]
+        }
+
+  defstruct headings: []
+end

packages/elixir/lib/kreuzberg/heading_level.ex

@@ -0,0 +1,17 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.HeadingLevel do
+  @moduledoc "A single heading in the hierarchy."
+
+  @typedoc "A single heading in the hierarchy."
+  @type t :: %__MODULE__{
+          level: non_neg_integer(),
+          text: String.t() | nil
+        }
+
+  defstruct level: 0,
+            text: nil
+end

packages/elixir/lib/kreuzberg/hierarchical_block.ex

@@ -0,0 +1,26 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.HierarchicalBlock do
+  @moduledoc """
+  A text block with hierarchy level assignment.
+
+  Represents a block of text with semantic heading information extracted from
+  font size clustering and hierarchical analysis.
+  """
+
+  @typedoc "A text block with hierarchy level assignment."
+  @type t :: %__MODULE__{
+          text: String.t() | nil,
+          font_size: float(),
+          level: String.t() | nil,
+          bbox: [float()] | nil
+        }
+
+  defstruct text: nil,
+            font_size: 0.0,
+            level: nil,
+            bbox: nil
+end

packages/elixir/lib/kreuzberg/hierarchy_config.ex

@@ -0,0 +1,38 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.HierarchyConfig do
+  @moduledoc """
+  Hierarchy extraction configuration for PDF text structure analysis.
+
+  Enables extraction of document hierarchy levels (H1-H6) based on font size
+  clustering and semantic analysis. When enabled, hierarchical blocks are
+  included in page content.
+  """
+
+  @typedoc "Hierarchy extraction configuration for PDF text structure analysis."
+  @type t :: %__MODULE__{
+          enabled: boolean(),
+          k_clusters: non_neg_integer(),
+          include_bbox: boolean(),
+          ocr_coverage_threshold: float() | nil
+        }
+
+  defstruct enabled: true,
+            k_clusters: 3,
+            include_bbox: true,
+            ocr_coverage_threshold: nil
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/html_metadata.ex

@@ -0,0 +1,59 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.HtmlMetadata do
+  @moduledoc """
+  HTML metadata extracted from HTML documents.
+
+  Includes document-level metadata, Open Graph data, Twitter Card metadata,
+  and extracted structural elements (headers, links, images, structured data).
+  """
+
+  @typedoc "HTML metadata extracted from HTML documents."
+  @type t :: %__MODULE__{
+          title: String.t() | nil,
+          description: String.t() | nil,
+          keywords: [String.t()],
+          author: String.t() | nil,
+          canonical_url: String.t() | nil,
+          base_href: String.t() | nil,
+          language: String.t() | nil,
+          text_direction: String.t() | nil | nil,
+          open_graph: map(),
+          twitter_card: map(),
+          meta_tags: map(),
+          headers: [map()],
+          links: [map()],
+          images: [map()],
+          structured_data: [map()]
+        }
+
+  defstruct title: nil,
+            description: nil,
+            keywords: [],
+            author: nil,
+            canonical_url: nil,
+            base_href: nil,
+            language: nil,
+            text_direction: nil,
+            open_graph: %{},
+            twitter_card: %{},
+            meta_tags: %{},
+            headers: [],
+            links: [],
+            images: [],
+            structured_data: []
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/html_output_config.ex

@@ -0,0 +1,53 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.HtmlOutputConfig do
+  @moduledoc """
+  Configuration for styled HTML output.
+
+  When set on [`ExtractionConfig::html_output`] alongside
+  `output_format = OutputFormat::Html`, the pipeline builds a
+  [`StyledHtmlRenderer`](crate::rendering::StyledHtmlRenderer) instead of
+  the plain comrak-based renderer.
+
+  # Example
+
+  ```rust
+  use kreuzberg::core::config::{HtmlOutputConfig, HtmlTheme};
+
+  let config = HtmlOutputConfig {
+      theme: HtmlTheme::GitHub,
+      css: Some(".kb-p { font-size: 1.1rem; }".to_string()),
+      ..Default::default()
+  };
+  ```
+  """
+
+  @typedoc "Configuration for styled HTML output."
+  @type t :: %__MODULE__{
+          css: String.t() | nil,
+          css_file: String.t() | nil,
+          theme: String.t() | nil,
+          class_prefix: String.t() | nil,
+          embed_css: boolean()
+        }
+
+  defstruct css: nil,
+            css_file: nil,
+            theme: :unstyled,
+            class_prefix: nil,
+            embed_css: true
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/html_theme.ex

@@ -0,0 +1,33 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.HtmlTheme do
+  @moduledoc "Built-in HTML theme selection."
+
+  @typedoc "Built-in HTML theme selection."
+  @type t :: :default | :git_hub | :dark | :light | :unstyled
+
+  @default :default
+  @git_hub :git_hub
+  @dark :dark
+  @light :light
+  @unstyled :unstyled
+
+  @doc "Sensible defaults: system font stack, neutral colours, readable line measure. CSS custom properties (`--kb-*`) are all defined so user CSS can override individual values."
+  @spec default() :: t()
+  def default, do: @default
+  @doc "GitHub Markdown-inspired palette and spacing."
+  @spec git_hub() :: t()
+  def git_hub, do: @git_hub
+  @doc "Dark background, light text."
+  @spec dark() :: t()
+  def dark, do: @dark
+  @doc "Minimal light theme with generous whitespace."
+  @spec light() :: t()
+  def light, do: @light
+  @doc "No built-in stylesheet emitted. CSS custom properties are still defined on `:root` so user stylesheets can reference `var(--kb-*)` tokens."
+  @spec unstyled() :: t()
+  def unstyled, do: @unstyled
+end

packages/elixir/lib/kreuzberg/image_extraction_config.ex

@@ -0,0 +1,50 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ImageExtractionConfig do
+  @moduledoc "Image extraction configuration."
+
+  @typedoc "Image extraction configuration."
+  @type t :: %__MODULE__{
+          extract_images: boolean(),
+          target_dpi: integer(),
+          max_image_dimension: integer(),
+          inject_placeholders: boolean(),
+          auto_adjust_dpi: boolean(),
+          min_dpi: integer(),
+          max_dpi: integer(),
+          max_images_per_page: non_neg_integer() | nil,
+          classify: boolean(),
+          include_page_rasters: boolean(),
+          run_ocr_on_images: boolean(),
+          ocr_text_only: boolean(),
+          append_ocr_text: boolean()
+        }
+
+  defstruct extract_images: true,
+            target_dpi: 300,
+            max_image_dimension: 4_096,
+            inject_placeholders: true,
+            auto_adjust_dpi: true,
+            min_dpi: 72,
+            max_dpi: 600,
+            max_images_per_page: nil,
+            classify: true,
+            include_page_rasters: false,
+            run_ocr_on_images: true,
+            ocr_text_only: false,
+            append_ocr_text: false
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/image_kind.ex

@@ -0,0 +1,73 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ImageKind do
+  @moduledoc "Heuristic classification of what an image likely depicts."
+
+  @typedoc "Heuristic classification of what an image likely depicts."
+  @type t ::
+          :photograph
+          | :diagram
+          | :chart
+          | :drawing
+          | :text_block
+          | :decoration
+          | :logo
+          | :icon
+          | :tile_fragment
+          | :mask
+          | :page_raster
+          | :unknown
+
+  @photograph :photograph
+  @diagram :diagram
+  @chart :chart
+  @drawing :drawing
+  @text_block :text_block
+  @decoration :decoration
+  @logo :logo
+  @icon :icon
+  @tile_fragment :tile_fragment
+  @mask :mask
+  @page_raster :page_raster
+  @unknown :unknown
+
+  @doc "Photographic image (natural scene, photograph)"
+  @spec photograph() :: t()
+  def photograph, do: @photograph
+  @doc "Technical or schematic diagram"
+  @spec diagram() :: t()
+  def diagram, do: @diagram
+  @doc "Chart, graph, or plot"
+  @spec chart() :: t()
+  def chart, do: @chart
+  @doc "Freehand or technical drawing"
+  @spec drawing() :: t()
+  def drawing, do: @drawing
+  @doc "Text-heavy image (scanned text, document)"
+  @spec text_block() :: t()
+  def text_block, do: @text_block
+  @doc "Decorative element or border"
+  @spec decoration() :: t()
+  def decoration, do: @decoration
+  @doc "Logo or brand mark"
+  @spec logo() :: t()
+  def logo, do: @logo
+  @doc "Small icon"
+  @spec icon() :: t()
+  def icon, do: @icon
+  @doc "Fragment of a larger tiled image (tile of a technical drawing)"
+  @spec tile_fragment() :: t()
+  def tile_fragment, do: @tile_fragment
+  @doc "Mask or transparency map"
+  @spec mask() :: t()
+  def mask, do: @mask
+  @doc "Full-page render produced during OCR preprocessing; used as a citation thumbnail."
+  @spec page_raster() :: t()
+  def page_raster, do: @page_raster
+  @doc "Could not classify with reasonable confidence"
+  @spec unknown() :: t()
+  def unknown, do: @unknown
+end

packages/elixir/lib/kreuzberg/image_metadata.ex

@@ -0,0 +1,36 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ImageMetadata do
+  @moduledoc """
+  Image metadata extracted from image files.
+
+  Includes dimensions, format, and EXIF data.
+  """
+
+  @typedoc "Image metadata extracted from image files."
+  @type t :: %__MODULE__{
+          width: non_neg_integer(),
+          height: non_neg_integer(),
+          format: String.t() | nil,
+          exif: map()
+        }
+
+  defstruct width: 0,
+            height: 0,
+            format: nil,
+            exif: %{}
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/image_metadata_type.ex

@@ -0,0 +1,25 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ImageMetadataType do
+  @moduledoc "Image element metadata."
+
+  @typedoc "Image element metadata."
+  @type t :: %__MODULE__{
+          src: String.t() | nil,
+          alt: String.t() | nil,
+          title: String.t() | nil,
+          dimensions: [non_neg_integer()] | nil,
+          image_type: String.t() | nil,
+          attributes: [[String.t()]]
+        }
+
+  defstruct src: nil,
+            alt: nil,
+            title: nil,
+            dimensions: nil,
+            image_type: :data_uri,
+            attributes: []
+end

packages/elixir/lib/kreuzberg/image_preprocessing_config.ex

@@ -0,0 +1,44 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ImagePreprocessingConfig do
+  @moduledoc """
+  Image preprocessing configuration for OCR.
+
+  These settings control how images are preprocessed before OCR to improve
+  text recognition quality. Different preprocessing strategies work better
+  for different document types.
+  """
+
+  @typedoc "Image preprocessing configuration for OCR."
+  @type t :: %__MODULE__{
+          target_dpi: integer(),
+          auto_rotate: boolean(),
+          deskew: boolean(),
+          denoise: boolean(),
+          contrast_enhance: boolean(),
+          binarization_method: String.t() | nil,
+          invert_colors: boolean()
+        }
+
+  defstruct target_dpi: 300,
+            auto_rotate: true,
+            deskew: true,
+            denoise: false,
+            contrast_enhance: false,
+            binarization_method: "otsu",
+            invert_colors: false
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/image_preprocessing_metadata.ex

@@ -0,0 +1,42 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ImagePreprocessingMetadata do
+  @moduledoc """
+  Image preprocessing metadata.
+
+  Tracks the transformations applied to an image during OCR preprocessing,
+  including DPI normalization, resizing, and resampling.
+  """
+
+  @typedoc "Image preprocessing metadata."
+  @type t :: %__MODULE__{
+          original_dimensions: [non_neg_integer()],
+          original_dpi: [float()],
+          target_dpi: integer(),
+          scale_factor: float(),
+          auto_adjusted: boolean(),
+          final_dpi: integer(),
+          new_dimensions: [non_neg_integer()] | nil,
+          resample_method: String.t() | nil,
+          dimension_clamped: boolean(),
+          calculated_dpi: integer() | nil,
+          skipped_resize: boolean(),
+          resize_error: String.t() | nil
+        }
+
+  defstruct original_dimensions: [],
+            original_dpi: [],
+            target_dpi: 0,
+            scale_factor: 0.0,
+            auto_adjusted: false,
+            final_dpi: 0,
+            new_dimensions: nil,
+            resample_method: nil,
+            dimension_clamped: false,
+            calculated_dpi: nil,
+            skipped_resize: false,
+            resize_error: nil
+end

packages/elixir/lib/kreuzberg/image_type.ex

@@ -0,0 +1,29 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.ImageType do
+  @moduledoc "Image type classification."
+
+  @typedoc "Image type classification."
+  @type t :: :"data-uri" | :"inline-svg" | :external | :relative
+
+  @data_uri :"data-uri"
+  @inline_svg :"inline-svg"
+  @external :external
+  @relative :relative
+
+  @doc "Data URI image"
+  @spec data_uri() :: t()
+  def data_uri, do: @data_uri
+  @doc "Inline SVG"
+  @spec inline_svg() :: t()
+  def inline_svg, do: @inline_svg
+  @doc "External image URL"
+  @spec external() :: t()
+  def external, do: @external
+  @doc "Relative path image"
+  @spec relative() :: t()
+  def relative, do: @relative
+end

packages/elixir/lib/kreuzberg/inline_element.ex

@@ -0,0 +1,25 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.InlineElement do
+  @moduledoc """
+  Inline element within a block.
+
+  Represents text with formatting, links, images, etc.
+  """
+
+  @typedoc "Inline element within a block."
+  @type t :: %__MODULE__{
+          element_type: String.t() | nil,
+          content: String.t() | nil,
+          attributes: String.t() | nil,
+          metadata: map() | nil
+        }
+
+  defstruct element_type: :text,
+            content: nil,
+            attributes: nil,
+            metadata: nil
+end

packages/elixir/lib/kreuzberg/inline_type.ex

@@ -0,0 +1,77 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.InlineType do
+  @moduledoc "Types of inline elements in Djot."
+
+  @typedoc "Types of inline elements in Djot."
+  @type t ::
+          :text
+          | :strong
+          | :emphasis
+          | :highlight
+          | :subscript
+          | :superscript
+          | :insert
+          | :delete
+          | :code
+          | :link
+          | :image
+          | :span
+          | :math
+          | :raw_inline
+          | :footnote_ref
+          | :symbol
+
+  @text :text
+  @strong :strong
+  @emphasis :emphasis
+  @highlight :highlight
+  @subscript :subscript
+  @superscript :superscript
+  @insert :insert
+  @delete :delete
+  @code :code
+  @link :link
+  @image :image
+  @span :span
+  @math :math
+  @raw_inline :raw_inline
+  @footnote_ref :footnote_ref
+  @symbol :symbol
+
+  @spec text() :: t()
+  def text, do: @text
+  @spec strong() :: t()
+  def strong, do: @strong
+  @spec emphasis() :: t()
+  def emphasis, do: @emphasis
+  @spec highlight() :: t()
+  def highlight, do: @highlight
+  @spec subscript() :: t()
+  def subscript, do: @subscript
+  @spec superscript() :: t()
+  def superscript, do: @superscript
+  @spec insert() :: t()
+  def insert, do: @insert
+  @spec delete() :: t()
+  def delete, do: @delete
+  @spec code() :: t()
+  def code, do: @code
+  @spec link() :: t()
+  def link, do: @link
+  @spec image() :: t()
+  def image, do: @image
+  @spec span() :: t()
+  def span, do: @span
+  @spec math() :: t()
+  def math, do: @math
+  @spec raw_inline() :: t()
+  def raw_inline, do: @raw_inline
+  @spec footnote_ref() :: t()
+  def footnote_ref, do: @footnote_ref
+  @spec symbol() :: t()
+  def symbol, do: @symbol
+end

packages/elixir/lib/kreuzberg/jats_metadata.ex

@@ -0,0 +1,32 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.JatsMetadata do
+  @moduledoc "JATS (Journal Article Tag Suite) metadata."
+
+  @typedoc "JATS (Journal Article Tag Suite) metadata."
+  @type t :: %__MODULE__{
+          copyright: String.t() | nil,
+          license: String.t() | nil,
+          history_dates: map(),
+          contributor_roles: [map()]
+        }
+
+  defstruct copyright: nil,
+            license: nil,
+            history_dates: %{},
+            contributor_roles: []
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/keyword.ex

@@ -0,0 +1,21 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.Keyword do
+  @moduledoc "Extracted keyword with metadata."
+
+  @typedoc "Extracted keyword with metadata."
+  @type t :: %__MODULE__{
+          text: String.t() | nil,
+          score: float(),
+          algorithm: String.t() | nil,
+          positions: [non_neg_integer()] | nil
+        }
+
+  defstruct text: nil,
+            score: 0.0,
+            algorithm: :yake,
+            positions: nil
+end

packages/elixir/lib/kreuzberg/keyword_algorithm.ex

@@ -0,0 +1,21 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.KeywordAlgorithm do
+  @moduledoc "Keyword algorithm selection."
+
+  @typedoc "Keyword algorithm selection."
+  @type t :: :yake | :rake
+
+  @yake :yake
+  @rake :rake
+
+  @doc "YAKE (Yet Another Keyword Extractor) - statistical approach"
+  @spec yake() :: t()
+  def yake, do: @yake
+  @doc "RAKE (Rapid Automatic Keyword Extraction) - co-occurrence based"
+  @spec rake() :: t()
+  def rake, do: @rake
+end

packages/elixir/lib/kreuzberg/keyword_config.ex

@@ -0,0 +1,38 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.KeywordConfig do
+  @moduledoc "Keyword extraction configuration."
+
+  @typedoc "Keyword extraction configuration."
+  @type t :: %__MODULE__{
+          algorithm: String.t() | nil,
+          max_keywords: non_neg_integer(),
+          min_score: float(),
+          ngram_range: [non_neg_integer()],
+          language: String.t() | nil,
+          yake_params: map() | nil,
+          rake_params: map() | nil
+        }
+
+  defstruct algorithm: :yake,
+            max_keywords: 10,
+            min_score: 0,
+            ngram_range: [],
+            language: nil,
+            yake_params: nil,
+            rake_params: nil
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/language_detection_config.ex

@@ -0,0 +1,30 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.LanguageDetectionConfig do
+  @moduledoc "Language detection configuration."
+
+  @typedoc "Language detection configuration."
+  @type t :: %__MODULE__{
+          enabled: boolean(),
+          min_confidence: float(),
+          detect_multiple: boolean()
+        }
+
+  defstruct enabled: true,
+            min_confidence: 0.8,
+            detect_multiple: false
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end

packages/elixir/lib/kreuzberg/layout_class.ex

@@ -0,0 +1,89 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.LayoutClass do
+  @moduledoc """
+  The 17 canonical document layout classes.
+
+  All model backends (RT-DETR, YOLO, etc.) map their native class IDs
+  to this shared set. Models with fewer classes (DocLayNet: 11, PubLayNet: 5)
+  map to the closest equivalent.
+
+  Wire format is snake_case in all serializers (JSON, TOML, YAML).
+  """
+
+  @typedoc "The 17 canonical document layout classes."
+  @type t ::
+          :caption
+          | :footnote
+          | :formula
+          | :list_item
+          | :page_footer
+          | :page_header
+          | :picture
+          | :section_header
+          | :table
+          | :text
+          | :title
+          | :document_index
+          | :code
+          | :checkbox_selected
+          | :checkbox_unselected
+          | :form
+          | :key_value_region
+
+  @caption :caption
+  @footnote :footnote
+  @formula :formula
+  @list_item :list_item
+  @page_footer :page_footer
+  @page_header :page_header
+  @picture :picture
+  @section_header :section_header
+  @table :table
+  @text :text
+  @title :title
+  @document_index :document_index
+  @code :code
+  @checkbox_selected :checkbox_selected
+  @checkbox_unselected :checkbox_unselected
+  @form :form
+  @key_value_region :key_value_region
+
+  @spec caption() :: t()
+  def caption, do: @caption
+  @spec footnote() :: t()
+  def footnote, do: @footnote
+  @spec formula() :: t()
+  def formula, do: @formula
+  @spec list_item() :: t()
+  def list_item, do: @list_item
+  @spec page_footer() :: t()
+  def page_footer, do: @page_footer
+  @spec page_header() :: t()
+  def page_header, do: @page_header
+  @spec picture() :: t()
+  def picture, do: @picture
+  @spec section_header() :: t()
+  def section_header, do: @section_header
+  @spec table() :: t()
+  def table, do: @table
+  @spec text() :: t()
+  def text, do: @text
+  @spec title() :: t()
+  def title, do: @title
+  @spec document_index() :: t()
+  def document_index, do: @document_index
+  @spec code() :: t()
+  def code, do: @code
+  @spec checkbox_selected() :: t()
+  def checkbox_selected, do: @checkbox_selected
+  @spec checkbox_unselected() :: t()
+  def checkbox_unselected, do: @checkbox_unselected
+  @spec form() :: t()
+  def form, do: @form
+  @spec key_value_region() :: t()
+  def key_value_region, do: @key_value_region
+end

packages/elixir/lib/kreuzberg/layout_detection.ex

@@ -0,0 +1,19 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.LayoutDetection do
+  @moduledoc "A single layout detection result."
+
+  @typedoc "A single layout detection result."
+  @type t :: %__MODULE__{
+          class_name: String.t() | nil,
+          confidence: float(),
+          bbox: map()
+        }
+
+  defstruct class_name: :caption,
+            confidence: 0.0,
+            bbox: nil
+end

packages/elixir/lib/kreuzberg/layout_detection_config.ex

@@ -0,0 +1,38 @@
+# This file is auto-generated by alef — DO NOT EDIT.
+# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
+# To regenerate: alef generate
+# To verify freshness: alef verify --exit-code
+# Issues & docs: https://github.com/kreuzberg-dev/alef
+defmodule Kreuzberg.LayoutDetectionConfig do
+  @moduledoc """
+  Layout detection configuration.
+
+  Controls layout detection behavior in the extraction pipeline.
+  When set on [`ExtractionConfig`](super::ExtractionConfig), layout detection
+  is enabled for PDF extraction.
+  """
+
+  @typedoc "Layout detection configuration."
+  @type t :: %__MODULE__{
+          confidence_threshold: float() | nil,
+          apply_heuristics: boolean(),
+          table_model: String.t() | nil,
+          acceleration: map() | nil
+        }
+
+  defstruct confidence_threshold: nil,
+            apply_heuristics: true,
+            table_model: :tatr,
+            acceleration: nil
+
+  defimpl Jason.Encoder do
+    @doc false
+    def encode(value, opts) do
+      value
+      |> Map.from_struct()
+      |> Enum.reject(fn {_k, v} -> v == nil end)
+      |> Enum.into(%{})
+      |> Jason.Encoder.encode(opts)
+    end
+  end
+end