Nomad changes

e2e/swift_e2e/Package.swift

@@ -0,0 +1,19 @@
+// swift-tools-version: 6.0
+import PackageDescription
+
+let package = Package(
+    name: "E2eSwift",
+    platforms: [
+        .macOS(.v13),
+        .iOS(.v16),
+    ],
+    dependencies: [
+        .package(path: "../../packages/swift"),
+    ],
+    targets: [
+        .testTarget(
+            name: "KreuzbergE2ETests",
+            dependencies: [.product(name: "Kreuzberg", package: "swift")]
+        ),
+    ]
+)

e2e/swift_e2e/Tests/KreuzbergE2ETests/AsyncTests.swift

@@ -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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: async.
+final class AsyncTests: XCTestCase {
+    override class func setUp() {
+        super.setUp()
+        let _testDocs = URL(fileURLWithPath: #filePath)
+            .deletingLastPathComponent() // <Module>Tests/
+            .deletingLastPathComponent() // Tests/
+            .deletingLastPathComponent() // swift/
+            .deletingLastPathComponent() // packages/
+            .deletingLastPathComponent() // <repo root>
+            .appendingPathComponent("test_documents")
+        if FileManager.default.fileExists(atPath: _testDocs.path) {
+            FileManager.default.changeCurrentDirectoryPath(_testDocs.path)
+        }
+    }
+
+    func testAsyncExtractBytes() throws {
+        // Async extract_bytes call on PDF document
+        let result = try Kreuzberg.extractBytes("pdf/fake_memo.pdf", "application/pdf", "{}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "application/pdf")
+        XCTAssertGreaterThanOrEqual(result.content.count, 50)
+    }
+
+    func testAsyncExtractBytesEmptyMime() throws {
+        // extract_bytes empty MIME async
+        do {
+            _ = try Kreuzberg.extractBytes("text/plain.txt", "", "{}")
+            XCTFail("expected to throw")
+        } catch {
+            // success
+        }
+    }
+
+    func testAsyncExtractBytesInvalidMime() throws {
+        // extract_bytes unsupported MIME async
+        do {
+            _ = try Kreuzberg.extractBytes("text/plain.txt", "application/x-nonexistent", "{}")
+            XCTFail("expected to throw")
+        } catch {
+            // success
+        }
+    }
+
+}

e2e/swift_e2e/Tests/KreuzbergE2ETests/BatchTests.swift

@@ -0,0 +1,93 @@
+// 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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: batch.
+final class BatchTests: XCTestCase {
+    func testBatchBytesInvalidMime() throws {
+        // batch_extract_bytes_sync invalid MIME
+        let _item_itemsArray_0 = try Kreuzberg.batchBytesItemFromJson("{\"content\":[72,101,108,108,111],\"mime_type\":\"application/x-nonexistent\"}")
+        let itemsArray = [_item_itemsArray_0]
+        let result = try Kreuzberg.batchExtractBytesSync(items: itemsArray)
+    }
+
+    func testBatchExtractBytesHappy() async throws {
+        // batch_extract_bytes: happy path with mixed inputs
+        let _item_itemsArray_0 = try Kreuzberg.batchBytesItemFromJson("{\"content\":[72,101,108,108,111,44,32,119,111,114,108,100,33],\"mime_type\":\"text/plain\"}")
+        let _item_itemsArray_1 = try Kreuzberg.batchBytesItemFromJson("{\"content\":[60,104,116,109,108,62,60,98,111,100,121,62,84,101,115,116,60,47,98,111,100,121,62,60,47,104,116,109,108,62],\"mime_type\":\"text/html\"}")
+        let itemsArray = [_item_itemsArray_0, _item_itemsArray_1]
+        let result = try await Kreuzberg.batchExtractBytes(items: itemsArray)
+        XCTAssertGreaterThanOrEqual(result.count, 1)
+    }
+
+    func testBatchExtractBytesMixedFormat() async throws {
+        // batch_extract_bytes: handles unsupported MIME gracefully
+        let _item_itemsArray_0 = try Kreuzberg.batchBytesItemFromJson("{\"content\":[80,68,70,32,112,108,97,99,101,104,111,108,100,101,114],\"mime_type\":\"application/x-unknown\"}")
+        let itemsArray = [_item_itemsArray_0]
+        let result = try await Kreuzberg.batchExtractBytes(items: itemsArray)
+    }
+
+    func testBatchExtractBytesSyncEmptyList() throws {
+        // batch_extract_bytes_sync: empty batch
+        let result = try Kreuzberg.batchExtractBytesSync(items: [])
+        XCTAssertEqual(result.count, 0)
+    }
+
+    func testBatchExtractBytesSyncInvalidMime() throws {
+        // batch_extract_bytes_sync: unsupported MIME
+        let _item_itemsArray_0 = try Kreuzberg.batchBytesItemFromJson("{\"content\":[100,97,116,97],\"mime_type\":\"application/x-unknown\"}")
+        let itemsArray = [_item_itemsArray_0]
+        let result = try Kreuzberg.batchExtractBytesSync(items: itemsArray)
+    }
+
+    func testBatchFileAsyncBasic() async throws {
+        // Extract text from multiple files asynchronously
+        let _item_pathsArray_0 = try Kreuzberg.batchFileItemFromJson("{\"path\":\"pdf/fake_memo.pdf\"}")
+        let _item_pathsArray_1 = try Kreuzberg.batchFileItemFromJson("{\"path\":\"text/fake_text.txt\"}")
+        let pathsArray = [_item_pathsArray_0, _item_pathsArray_1]
+        let result = try await Kreuzberg.batchExtractFiles(paths: pathsArray)
+    }
+
+    func testBatchFileAsyncNotFound() async throws {
+        // batch_extract_file async nonexistent
+        let _item_pathsArray_0 = try Kreuzberg.batchFileItemFromJson("{\"path\":\"/nonexistent/a.pdf\"}")
+        let pathsArray = [_item_pathsArray_0]
+        let result = try await Kreuzberg.batchExtractFiles(paths: pathsArray)
+    }
+
+    func testBatchFileNotFound() throws {
+        // batch_extract_file_sync nonexistent
+        let _item_pathsArray_0 = try Kreuzberg.batchFileItemFromJson("{\"path\":\"/nonexistent/a.pdf\"}")
+        let _item_pathsArray_1 = try Kreuzberg.batchFileItemFromJson("{\"path\":\"/nonexistent/b.txt\"}")
+        let pathsArray = [_item_pathsArray_0, _item_pathsArray_1]
+        let result = try Kreuzberg.batchExtractFilesSync(paths: pathsArray)
+    }
+
+    func testBatchFilePartial() throws {
+        // batch_extract_file_sync mixed
+        let _item_pathsArray_0 = try Kreuzberg.batchFileItemFromJson("{\"path\":\"text/plain.txt\"}")
+        let _item_pathsArray_1 = try Kreuzberg.batchFileItemFromJson("{\"path\":\"/nonexistent/missing.pdf\"}")
+        let pathsArray = [_item_pathsArray_0, _item_pathsArray_1]
+        let result = try Kreuzberg.batchExtractFilesSync(paths: pathsArray)
+    }
+
+    func testBatchFileSyncBasic() throws {
+        // Extract text from multiple files synchronously
+        let _item_pathsArray_0 = try Kreuzberg.batchFileItemFromJson("{\"path\":\"pdf/fake_memo.pdf\"}")
+        let _item_pathsArray_1 = try Kreuzberg.batchFileItemFromJson("{\"path\":\"text/fake_text.txt\"}")
+        let pathsArray = [_item_pathsArray_0, _item_pathsArray_1]
+        let result = try Kreuzberg.batchExtractFilesSync(paths: pathsArray)
+    }
+
+}

e2e/swift_e2e/Tests/KreuzbergE2ETests/CodeTests.swift

@@ -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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: code.
+final class CodeTests: XCTestCase {
+    override class func setUp() {
+        super.setUp()
+        let _testDocs = URL(fileURLWithPath: #filePath)
+            .deletingLastPathComponent() // <Module>Tests/
+            .deletingLastPathComponent() // Tests/
+            .deletingLastPathComponent() // swift/
+            .deletingLastPathComponent() // packages/
+            .deletingLastPathComponent() // <repo root>
+            .appendingPathComponent("test_documents")
+        if FileManager.default.fileExists(atPath: _testDocs.path) {
+            FileManager.default.changeCurrentDirectoryPath(_testDocs.path)
+        }
+    }
+
+    func testCodeShebangDetection() throws {
+        // Test language detection from shebang line via bytes input
+        let result = try Kreuzberg.extractFileSync("code/script.sh", "text/x-source-code", "{}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "text/x-source-code")
+        XCTAssertGreaterThanOrEqual(result.content.count, 10)
+        XCTAssertTrue(result.content.contains("build"), "expected to contain: \("build")")
+        XCTAssertTrue(result.content.contains("clean"), "expected to contain: \("clean")")
+    }
+
+}

e2e/swift_e2e/Tests/KreuzbergE2ETests/ContractTests.swift

@@ -0,0 +1,169 @@
+// 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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: contract.
+final class ContractTests: XCTestCase {
+    override class func setUp() {
+        super.setUp()
+        let _testDocs = URL(fileURLWithPath: #filePath)
+            .deletingLastPathComponent() // <Module>Tests/
+            .deletingLastPathComponent() // Tests/
+            .deletingLastPathComponent() // swift/
+            .deletingLastPathComponent() // packages/
+            .deletingLastPathComponent() // <repo root>
+            .appendingPathComponent("test_documents")
+        if FileManager.default.fileExists(atPath: _testDocs.path) {
+            FileManager.default.changeCurrentDirectoryPath(_testDocs.path)
+        }
+    }
+
+    func testApiBatchBytesAsync() throws {
+        // Tests async batch bytes extraction API (batch_extract_bytes)
+        let result = try Kreuzberg.extractFile("pdf/fake_memo.pdf", nil, "{}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "application/pdf")
+        XCTAssertGreaterThanOrEqual(result.content.count, 10)
+        XCTAssertTrue(result.content.contains("May 5, 2023") || result.content.contains("Mallori"), "expected to contain at least one of the specified values")
+    }
+
+    func testApiBatchBytesWithConfigsAsync() throws {
+        // Tests async batch bytes extraction with per-file configs (batch_extract_bytes with file_configs parameter)
+        let result = try Kreuzberg.extractFile("pdf/fake_memo.pdf", nil, "{\"output_format\":\"markdown\"}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "application/pdf")
+        XCTAssertGreaterThanOrEqual(result.content.count, 10)
+        // skipped: field 'metadata.output_format' not available on result type
+    }
+
+    func testApiBatchFileAsync() throws {
+        // Tests async batch file extraction API (batch_extract_file)
+        let result = try Kreuzberg.extractFile("pdf/fake_memo.pdf", nil, "{}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "application/pdf")
+        XCTAssertGreaterThanOrEqual(result.content.count, 10)
+        XCTAssertTrue(result.content.contains("May 5, 2023") || result.content.contains("Mallori"), "expected to contain at least one of the specified values")
+    }
+
+    func testApiBatchFileWithConfigsAsync() throws {
+        // Tests async batch file extraction with per-file configs (batch_extract_files with file_configs parameter)
+        let result = try Kreuzberg.extractFile("pdf/fake_memo.pdf", nil, "{\"output_format\":\"markdown\"}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "application/pdf")
+        XCTAssertGreaterThanOrEqual(result.content.count, 10)
+        // skipped: field 'metadata.output_format' not available on result type
+    }
+
+    func testApiExtractBytesAsync() throws {
+        // Tests async bytes extraction API (extract_bytes)
+        let result = try Kreuzberg.extractFile("pdf/fake_memo.pdf", nil, "{}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "application/pdf")
+        XCTAssertGreaterThanOrEqual(result.content.count, 10)
+        XCTAssertTrue(result.content.contains("May 5, 2023") || result.content.contains("Mallori"), "expected to contain at least one of the specified values")
+    }
+
+    func testApiExtractFileAsync() async throws {
+        // Tests async file extraction API (extract_file)
+        let result = try await Kreuzberg.extractFile("pdf/fake_memo.pdf", nil, "{}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "application/pdf")
+        XCTAssertGreaterThanOrEqual(result.content.count, 10)
+        XCTAssertTrue(result.content.contains("May 5, 2023") || result.content.contains("Mallori"), "expected to contain at least one of the specified values")
+    }
+
+    func testConfigChunkingPrependHeadingContext() throws {
+        // Tests markdown chunker prepends heading hierarchy to chunk content
+        let result = try Kreuzberg.extractFileSync("markdown/extraction_test.md", nil, "{\"chunking\":{\"chunker_type\":\"markdown\",\"max_chars\":300,\"max_overlap\":50,\"prepend_heading_context\":true}}")
+        XCTAssertGreaterThanOrEqual(result.content.count, 10)
+        // skipped: field 'chunks' not available on result type
+        // skipped: field 'chunks_have_content' not available on result type
+        // skipped: field 'chunks_have_heading_context' not available on result type
+        // skipped: field 'first_chunk_starts_with_heading' not available on result type
+    }
+
+    func testConfigDocumentStructureWithHeadings() throws {
+        // Tests document structure with DOCX heading-driven nesting
+        let result = try Kreuzberg.extractFileSync("docx/fake.docx", nil, "{\"include_document_structure\":true}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "application/vnd.openxmlformats-officedocument.wordprocessingml.document")
+        // skipped: field 'document' not available on result type
+        // skipped: field 'document.nodes' not available on result type
+    }
+
+    func testConfigElementTypes() throws {
+        // Tests element-based result format with element type assertions on DOCX
+        let result = try Kreuzberg.extractFileSync("docx/unit_test_headers.docx", nil, "{\"result_format\":\"element_based\"}")
+        XCTAssertTrue(result.mimeType.contains("application/vnd.openxmlformats-officedocument.wordprocessingml.document"), "expected to contain at least one of the specified values")
+        // skipped: field 'elements' not available on result type
+    }
+
+    func testConfigExtractionTimeout() throws {
+        // Tests that extraction_timeout_secs config field is accepted and does not affect fast extractions
+        let result = try Kreuzberg.extractFileSync("pdf/fake_memo.pdf", nil, "{\"extraction_timeout_secs\":300}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "application/pdf")
+        XCTAssertGreaterThanOrEqual(result.content.count, 10)
+    }
+
+    func testConfigKeywords() throws {
+        // Tests keyword extraction via YAKE algorithm
+        let result = try Kreuzberg.extractFileSync("pdf/fake_memo.pdf", nil, "{\"keywords\":{\"algorithm\":\"yake\",\"max_keywords\":10}}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "application/pdf")
+        XCTAssertGreaterThanOrEqual(result.content.count, 10)
+        // skipped: field 'keywords' not available on result type
+        // skipped: field 'keywords' not available on result type
+    }
+
+    func testConfigPages() throws {
+        // Tests page extraction and page marker configuration
+        let result = try Kreuzberg.extractFileSync("pdf/fake_memo.pdf", nil, "{\"pages\":{\"extract_pages\":true,\"insert_page_markers\":true}}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "application/pdf")
+        XCTAssertGreaterThanOrEqual(result.content.count, 10)
+        XCTAssertTrue(result.content.contains("PAGE"), "expected to contain at least one of the specified values")
+    }
+
+    func testConfigQualityEnabled() throws {
+        // Tests quality scoring produces a score value in [0.0, 1.0]
+        let result = try Kreuzberg.extractFileSync("pdf/fake_memo.pdf", nil, "{\"enable_quality_processing\":true}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "application/pdf")
+        XCTAssertGreaterThanOrEqual(result.content.count, 10)
+        // skipped: field 'quality_score' not available on result type
+        // skipped: field 'quality_score' not available on result type
+        // skipped: field 'quality_score' not available on result type
+    }
+
+    func testConfigSecurityLimits() throws {
+        // Tests archive extraction with custom security limits
+        let result = try Kreuzberg.extractFileSync("archives/documents.zip", nil, "{\"security_limits\":{\"max_archive_size\":104857600,\"max_compression_ratio\":50,\"max_files_in_archive\":100}}")
+        XCTAssertTrue(result.mimeType.contains("application/zip") || result.mimeType.contains("application/x-zip-compressed"), "expected to contain at least one of the specified values")
+        XCTAssertGreaterThanOrEqual(result.content.count, 10)
+    }
+
+    func testConfigTreeSitter() throws {
+        // Tests tree-sitter configuration round-trip
+        let result = try Kreuzberg.extractFileSync("code/hello.py", nil, "{\"tree_sitter\":{\"groups\":[\"web\"],\"languages\":[\"python\",\"rust\"],\"process\":{\"comments\":false,\"diagnostics\":false,\"docstrings\":false,\"exports\":true,\"imports\":true,\"structure\":true,\"symbols\":false}}}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "text/x-source-code")
+        XCTAssertGreaterThanOrEqual(result.content.count, 5)
+    }
+
+    func testOutputFormatBytesMarkdown() throws {
+        // Tests markdown output format via bytes extraction API
+        let result = try Kreuzberg.extractBytesSync("pdf/fake_memo.pdf", "application/pdf", "{\"output_format\":\"markdown\"}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "application/pdf")
+        XCTAssertGreaterThanOrEqual(result.content.count, 10)
+        // skipped: field 'metadata.output_format' not available on result type
+    }
+
+    func testOutputFormatMarkdown() throws {
+        // Tests Markdown output format
+        let result = try Kreuzberg.extractFileSync("pdf/fake_memo.pdf", nil, "{\"output_format\":\"markdown\"}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "application/pdf")
+        XCTAssertGreaterThanOrEqual(result.content.count, 10)
+        // skipped: field 'metadata.output_format' not available on result type
+    }
+
+}

e2e/swift_e2e/Tests/KreuzbergE2ETests/DetectionTests.swift

@@ -0,0 +1,57 @@
+// 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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: detection.
+final class DetectionTests: XCTestCase {
+    override class func setUp() {
+        super.setUp()
+        let _testDocs = URL(fileURLWithPath: #filePath)
+            .deletingLastPathComponent() // <Module>Tests/
+            .deletingLastPathComponent() // Tests/
+            .deletingLastPathComponent() // swift/
+            .deletingLastPathComponent() // packages/
+            .deletingLastPathComponent() // <repo root>
+            .appendingPathComponent("test_documents")
+        if FileManager.default.fileExists(atPath: _testDocs.path) {
+            FileManager.default.changeCurrentDirectoryPath(_testDocs.path)
+        }
+    }
+
+    func testDetectMimeBytesHtml() throws {
+        // Detect HTML MIME from bytes
+        let result = try Kreuzberg.detectMimeTypeFromBytes("html/html.html")
+    }
+
+    func testDetectMimeBytesPdf() throws {
+        // Detect PDF MIME type from bytes
+        let result = try Kreuzberg.detectMimeTypeFromBytes("pdf/fake_memo.pdf")
+    }
+
+    func testDetectMimeBytesPng() throws {
+        // Detect PNG MIME type from bytes
+        let result = try Kreuzberg.detectMimeTypeFromBytes("images/test_hello_world.png")
+    }
+
+    func testGetExtensionsUnknownMime() throws {
+        // get_extensions unknown MIME
+        do {
+            _ = try Kreuzberg.getExtensionsForMime(mimeType: "application/x-totally-unknown")
+            XCTFail("expected to throw")
+        } catch {
+            // success
+        }
+    }
+
+}

e2e/swift_e2e/Tests/KreuzbergE2ETests/DocumentExtractorManagementTests.swift

@@ -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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: document_extractor_management.
+final class DocumentExtractorManagementTests: XCTestCase {
+    func testDocumentExtractorsClear() throws {
+        // Clear all document extractors and verify list is empty
+        let result = try Kreuzberg.clearDocumentExtractors()
+    }
+
+    func testExtractorsList() throws {
+        // List all registered document extractors
+        let result = try Kreuzberg.listDocumentExtractors()
+    }
+
+}

e2e/swift_e2e/Tests/KreuzbergE2ETests/EmbedExtraTests.swift

@@ -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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: embed_extra.
+final class EmbedExtraTests: XCTestCase {
+    func testEmbedTextsBatch() throws {
+        // Batch embed texts
+        let configObj = try Kreuzberg.embeddingConfigFromJson("{\"model\":{\"name\":\"balanced\",\"type\":\"preset\"}}")
+        let result = try Kreuzberg.embedTexts(texts: ["Hello", "World"], config: configObj)
+    }
+
+}

e2e/swift_e2e/Tests/KreuzbergE2ETests/EmbeddingBackendManagementTests.swift

@@ -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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: embedding_backend_management.
+final class EmbeddingBackendManagementTests: XCTestCase {
+    func testEmbeddingBackendsClear() throws {
+        // Clear all embedding backends and verify list is empty
+        let result = try Kreuzberg.clearEmbeddingBackends()
+    }
+
+    func testEmbeddingBackendsList() throws {
+        // List all registered embedding backends
+        let result = try Kreuzberg.listEmbeddingBackends()
+    }
+
+}

e2e/swift_e2e/Tests/KreuzbergE2ETests/EmbeddingsTests.swift

@@ -0,0 +1,47 @@
+// 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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: embeddings.
+final class EmbeddingsTests: XCTestCase {
+    func testEmbedTextsDifferentPreset() throws {
+        // embed_texts: multilingual preset
+        let configObj = try Kreuzberg.embeddingConfigFromJson("{\"model\":{\"name\":\"multilingual\",\"type\":\"preset\"}}")
+        let result = try Kreuzberg.embedTexts(texts: ["Hello world", "Test"], config: configObj)
+        // skipped: field 'embeddings' not available on result type
+    }
+
+    func testGetEmbeddingPresetKnown() throws {
+        // get_embedding_preset: known preset
+        let result = try Kreuzberg.getEmbeddingPreset(name: "balanced")
+    }
+
+    func testGetEmbeddingPresetNominal() throws {
+        // get_embedding_preset: nominal case
+        let result = try Kreuzberg.getEmbeddingPreset(name: "balanced")
+    }
+
+    func testGetEmbeddingPresetUnknown() throws {
+        // get_embedding_preset: unknown preset fails
+        let result = try Kreuzberg.getEmbeddingPreset(name: "nonexistent-xyz")
+        XCTAssertNil(result, "expected nil value")
+    }
+
+    func testListEmbeddingPresetsSanity() throws {
+        // list_embedding_presets: returns at least one
+        let result = try Kreuzberg.listEmbeddingPresets()
+        XCTAssertFalse(result.isEmpty, "expected non-empty value")
+    }
+
+}

e2e/swift_e2e/Tests/KreuzbergE2ETests/ErrorTests.swift

@@ -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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: error.
+final class ErrorTests: XCTestCase {
+    override class func setUp() {
+        super.setUp()
+        let _testDocs = URL(fileURLWithPath: #filePath)
+            .deletingLastPathComponent() // <Module>Tests/
+            .deletingLastPathComponent() // Tests/
+            .deletingLastPathComponent() // swift/
+            .deletingLastPathComponent() // packages/
+            .deletingLastPathComponent() // <repo root>
+            .appendingPathComponent("test_documents")
+        if FileManager.default.fileExists(atPath: _testDocs.path) {
+            FileManager.default.changeCurrentDirectoryPath(_testDocs.path)
+        }
+    }
+
+    func testErrorEmptyBytes() throws {
+        // Graceful handling of empty bytes (should not error)
+        let result = try Kreuzberg.extractBytesSync("text/empty.txt", "text/plain", "{}")
+    }
+
+    func testErrorEmptyMime() throws {
+        // Error when extracting with empty MIME type
+        do {
+            _ = try Kreuzberg.extractBytesSync("text/plain.txt", "", "{}")
+            XCTFail("expected to throw")
+        } catch {
+            // success
+        }
+    }
+
+    func testErrorExtractBytesConflictingOcr() throws {
+        // extract_bytes force+disable OCR
+        do {
+            _ = try Kreuzberg.extractBytesSync("text/fake_text.txt", "text/plain", "{\"disable_ocr\":true,\"force_ocr\":true}")
+            XCTFail("expected to throw")
+        } catch {
+            // success
+        }
+    }
+
+    func testErrorInvalidMimeFormat() throws {
+        // Error when extracting with invalid MIME type format
+        do {
+            _ = try Kreuzberg.extractBytesSync("text/plain.txt", "not-a-mime", "{}")
+            XCTFail("expected to throw")
+        } catch {
+            // success
+        }
+    }
+
+    func testErrorUnsupportedMime() throws {
+        // Error when extracting with unsupported MIME type
+        do {
+            _ = try Kreuzberg.extractBytesSync("text/plain.txt", "application/x-nonexistent", "{}")
+            XCTFail("expected to throw")
+        } catch {
+            // success
+        }
+    }
+
+}

e2e/swift_e2e/Tests/KreuzbergE2ETests/FormatSpecificTests.swift

@@ -0,0 +1,62 @@
+// 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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: format_specific.
+final class FormatSpecificTests: XCTestCase {
+    override class func setUp() {
+        super.setUp()
+        let _testDocs = URL(fileURLWithPath: #filePath)
+            .deletingLastPathComponent() // <Module>Tests/
+            .deletingLastPathComponent() // Tests/
+            .deletingLastPathComponent() // swift/
+            .deletingLastPathComponent() // packages/
+            .deletingLastPathComponent() // <repo root>
+            .appendingPathComponent("test_documents")
+        if FileManager.default.fileExists(atPath: _testDocs.path) {
+            FileManager.default.changeCurrentDirectoryPath(_testDocs.path)
+        }
+    }
+
+    func testFormatDocxStandalone() throws {
+        // Standalone DOCX extraction using extract_bytes_sync
+        let result = try Kreuzberg.extractBytesSync("docx/fake.docx", "application/vnd.openxmlformats-officedocument.wordprocessingml.document", "{}")
+        XCTAssertGreaterThanOrEqual(result.content.count, 20)
+    }
+
+    func testFormatHwpxStandalone() throws {
+        // Standalone HWPX extraction using extract_bytes_sync
+        let result = try Kreuzberg.extractBytesSync("hwpx/simple.hwpx", "application/haansofthwpx", "{}")
+        XCTAssertGreaterThanOrEqual(result.content.count, 20)
+        XCTAssertTrue("Hello from HWPX".isEmpty || result.content.contains("Hello from HWPX"), "expected to contain: \("Hello from HWPX")")
+    }
+
+    func testFormatPdfText() throws {
+        // Standalone PDF text extraction using extract_bytes_sync
+        let result = try Kreuzberg.extractBytesSync("pdf/fake_memo.pdf", "application/pdf", "{}")
+        XCTAssertGreaterThanOrEqual(result.content.count, 50)
+        XCTAssertTrue(result.content.contains("Mallori") || result.content.contains("May"), "expected to contain at least one of the specified values")
+    }
+
+    func testFormatPptx() throws {
+        // PPTX presentation extraction using extract_file_sync
+        let result = try Kreuzberg.extractFileSync("pptx/simple.pptx", "application/vnd.openxmlformats-officedocument.presentationml.presentation", "{}")
+    }
+
+    func testFormatXlsx() throws {
+        // XLSX spreadsheet extraction using extract_file_sync
+        let result = try Kreuzberg.extractFileSync("xlsx/stanley_cups.xlsx", "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet", "{}")
+    }
+
+}

e2e/swift_e2e/Tests/KreuzbergE2ETests/MimeUtilitiesTests.swift

@@ -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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: mime_utilities.
+final class MimeUtilitiesTests: XCTestCase {
+    override class func setUp() {
+        super.setUp()
+        let _testDocs = URL(fileURLWithPath: #filePath)
+            .deletingLastPathComponent() // <Module>Tests/
+            .deletingLastPathComponent() // Tests/
+            .deletingLastPathComponent() // swift/
+            .deletingLastPathComponent() // packages/
+            .deletingLastPathComponent() // <repo root>
+            .appendingPathComponent("test_documents")
+        if FileManager.default.fileExists(atPath: _testDocs.path) {
+            FileManager.default.changeCurrentDirectoryPath(_testDocs.path)
+        }
+    }
+
+    func testMimeDetectBytes() throws {
+        // Detect MIME type from file bytes
+        let result = try Kreuzberg.detectMimeTypeFromBytes("pdf/fake_memo.pdf")
+        XCTAssertTrue("pdf".isEmpty || result.contains("pdf"), "expected to contain: \("pdf")")
+    }
+
+    func testMimeDetectImage() throws {
+        // Detect MIME type from PNG image bytes
+        let result = try Kreuzberg.detectMimeTypeFromBytes("images/test_hello_world.png")
+        XCTAssertTrue("png".isEmpty || result.contains("png"), "expected to contain: \("png")")
+    }
+
+    func testMimeGetExtensions() throws {
+        // Get file extensions for a MIME type
+        let result = try Kreuzberg.getExtensionsForMime(mimeType: "application/pdf")
+        XCTAssertTrue("pdf".isEmpty || result.contains("pdf"), "expected to contain: \("pdf")")
+    }
+
+}

e2e/swift_e2e/Tests/KreuzbergE2ETests/OcrBackendManagementTests.swift

@@ -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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: ocr_backend_management.
+final class OcrBackendManagementTests: XCTestCase {
+    func testOcrBackendsClear() throws {
+        // Clear all OCR backends and verify list is empty
+        let result = try Kreuzberg.clearOcrBackends()
+    }
+
+    func testOcrBackendsList() throws {
+        // List all registered OCR backends
+        let result = try Kreuzberg.listOcrBackends()
+    }
+
+    func testOcrBackendsUnregister() throws {
+        // Unregister nonexistent OCR backend gracefully
+        let result = try Kreuzberg.unregisterOcrBackend(name: "nonexistent-backend-xyz")
+    }
+
+}

e2e/swift_e2e/Tests/KreuzbergE2ETests/PdfTests.swift

@@ -0,0 +1,52 @@
+// 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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: pdf.
+final class PdfTests: XCTestCase {
+    override class func setUp() {
+        super.setUp()
+        let _testDocs = URL(fileURLWithPath: #filePath)
+            .deletingLastPathComponent() // <Module>Tests/
+            .deletingLastPathComponent() // Tests/
+            .deletingLastPathComponent() // swift/
+            .deletingLastPathComponent() // packages/
+            .deletingLastPathComponent() // <repo root>
+            .appendingPathComponent("test_documents")
+        if FileManager.default.fileExists(atPath: _testDocs.path) {
+            FileManager.default.changeCurrentDirectoryPath(_testDocs.path)
+        }
+    }
+
+    func testRenderPdfPageFirst() throws {
+        // render_pdf_page_to_png: first page
+        let pdfBytesData = try Data(contentsOf: URL(fileURLWithPath: "pdf/fake_memo.pdf"))
+        let pdfBytesBytes = Array(pdfBytesData)
+        let result = try Kreuzberg.renderPdfPageToPng(pdfBytes: pdfBytesBytes, pageIndex: 0, dpi: nil, password: nil)
+        XCTAssertGreaterThanOrEqual(result.count, 100)
+    }
+
+    func testRenderPdfPageOutOfRange() throws {
+        // render_pdf_page_to_png: page out of range
+        do {
+            let pdfBytesData = try Data(contentsOf: URL(fileURLWithPath: "pdf/fake_memo.pdf"))
+            let pdfBytesBytes = Array(pdfBytesData)
+            _ = try Kreuzberg.renderPdfPageToPng(pdfBytes: pdfBytesBytes, pageIndex: 999, dpi: nil, password: nil)
+            XCTFail("expected to throw")
+        } catch {
+            // success
+        }
+    }
+
+}

e2e/swift_e2e/Tests/KreuzbergE2ETests/PluginApiTests.swift

@@ -0,0 +1,114 @@
+// 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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: plugin_api.
+final class PluginApiTests: XCTestCase {
+    func testRegisterDocumentExtractorTraitBridge() throws {
+        // register_document_extractor: trait bridge
+        class TestStubRegisterDocumentExtractorTraitBridge: SwiftDocumentExtractorBridge {
+    var name: String { "register_document_extractor_trait_bridge" }
+    func extractBytes(content: Data, mimeType: String, config: ExtractionConfig) async throws -> InternalDocument { InternalDocument() }
+    func supportedMimeTypes() -> [String] { [] }
+}
+
+        let result = try Kreuzberg.registerDocumentExtractor(TestStubRegisterDocumentExtractorTraitBridge())
+        try? Kreuzberg.unregisterDocumentExtractor("swift-bridge-document_extractor")
+    }
+
+    func testRegisterEmbeddingBackendTraitBridge() throws {
+        // register_embedding_backend: trait bridge
+        class TestStubRegisterEmbeddingBackendTraitBridge: SwiftEmbeddingBackendBridge {
+    var name: String { "register_embedding_backend_trait_bridge" }
+    func dimensions() -> UInt { 0 }
+    func embed(texts: [String]) async throws -> [[Float]] { [] }
+}
+
+        let result = try Kreuzberg.registerEmbeddingBackend(TestStubRegisterEmbeddingBackendTraitBridge())
+        try? Kreuzberg.unregisterEmbeddingBackend("swift-bridge-embedding_backend")
+    }
+
+    func testRegisterOcrBackendTraitBridge() throws {
+        // register_ocr_backend: trait bridge
+        class TestStubRegisterOcrBackendTraitBridge: SwiftOcrBackendBridge {
+    var name: String { "register_ocr_backend_trait_bridge" }
+    func processImage(imageBytes: Data, config: OcrConfig) async throws -> ExtractionResult { ExtractionResult() }
+    func supportsLanguage(lang: String) -> Bool { false }
+    func backendType() -> OcrBackendType { OcrBackendType() }
+}
+
+        let result = try Kreuzberg.registerOcrBackend(TestStubRegisterOcrBackendTraitBridge())
+        try? Kreuzberg.unregisterOcrBackend("swift-bridge-ocr_backend")
+    }
+
+    func testRegisterPostProcessorTraitBridge() throws {
+        // register_post_processor: trait bridge
+        class TestStubRegisterPostProcessorTraitBridge: SwiftPostProcessorBridge {
+    var name: String { "register_post_processor_trait_bridge" }
+    func process(result: ExtractionResult, config: ExtractionConfig) async throws -> Void { () }
+    func processingStage() -> ProcessingStage { ProcessingStage() }
+}
+
+        let result = try Kreuzberg.registerPostProcessor(TestStubRegisterPostProcessorTraitBridge())
+        try? Kreuzberg.unregisterPostProcessor("swift-bridge-post_processor")
+    }
+
+    func testRegisterRendererTraitBridge() throws {
+        // register_renderer: trait bridge
+        class TestStubRegisterRendererTraitBridge: SwiftRendererBridge {
+    var name: String { "register_renderer_trait_bridge" }
+    func render(doc: InternalDocument) -> String { "" }
+}
+
+        let result = try Kreuzberg.registerRenderer(TestStubRegisterRendererTraitBridge())
+        try? Kreuzberg.unregisterRenderer("swift-bridge-renderer")
+    }
+
+    func testRegisterValidatorTraitBridge() throws {
+        // register_validator: trait bridge
+        class TestStubRegisterValidatorTraitBridge: SwiftValidatorBridge {
+    var name: String { "register_validator_trait_bridge" }
+    func validate(result: ExtractionResult, config: ExtractionConfig) async throws -> Void { () }
+}
+
+        let result = try Kreuzberg.registerValidator(TestStubRegisterValidatorTraitBridge())
+        try? Kreuzberg.unregisterValidator("swift-bridge-validator")
+    }
+
+    func testUnregisterDocumentExtractorAfterRegister() throws {
+        // unregister_document_extractor
+        let result = try Kreuzberg.unregisterDocumentExtractor(name: "test-extractor")
+    }
+
+    func testUnregisterEmbeddingBackendAfterRegister() throws {
+        // unregister_embedding_backend
+        let result = try Kreuzberg.unregisterEmbeddingBackend(name: "test-embedding-backend")
+    }
+
+    func testUnregisterPostProcessorAfterRegister() throws {
+        // unregister_post_processor
+        let result = try Kreuzberg.unregisterPostProcessor(name: "test-processor")
+    }
+
+    func testUnregisterRendererAfterRegister() throws {
+        // unregister_renderer
+        let result = try Kreuzberg.unregisterRenderer(name: "test-renderer")
+    }
+
+    func testUnregisterValidatorAfterRegister() throws {
+        // unregister_validator
+        let result = try Kreuzberg.unregisterValidator(name: "test-validator")
+    }
+
+}

e2e/swift_e2e/Tests/KreuzbergE2ETests/PostProcessorManagementTests.swift

@@ -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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: post_processor_management.
+final class PostProcessorManagementTests: XCTestCase {
+    func testPostProcessorsClear() throws {
+        // Clear all post-processors and verify list is empty
+        let result = try Kreuzberg.clearPostProcessors()
+    }
+
+    func testPostProcessorsList() throws {
+        // List all registered post-processors
+        let result = try Kreuzberg.listPostProcessors()
+    }
+
+}

e2e/swift_e2e/Tests/KreuzbergE2ETests/RegistryOperationsTests.swift

@@ -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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: registry_operations.
+final class RegistryOperationsTests: XCTestCase {
+    func testExtensionsDocx() throws {
+        // Get file extensions for DOCX MIME type
+        let result = try Kreuzberg.getExtensionsForMime(mimeType: "application/vnd.openxmlformats-officedocument.wordprocessingml.document")
+    }
+
+    func testExtensionsHtml() throws {
+        // Get file extensions for HTML MIME type
+        let result = try Kreuzberg.getExtensionsForMime(mimeType: "text/html")
+    }
+
+    func testExtensionsPdf() throws {
+        // Get file extensions for PDF MIME type
+        let result = try Kreuzberg.getExtensionsForMime(mimeType: "application/pdf")
+    }
+
+}

e2e/swift_e2e/Tests/KreuzbergE2ETests/RegistryTests.swift

@@ -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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: registry.
+final class RegistryTests: XCTestCase {
+    func testListDocumentExtractors() throws {
+        // List document extractors
+        let result = try Kreuzberg.listDocumentExtractors()
+    }
+
+    func testListEmbeddingBackends() throws {
+        // List embedding backends
+        let result = try Kreuzberg.listEmbeddingBackends()
+    }
+
+    func testListOcrBackends() throws {
+        // List OCR backends
+        let result = try Kreuzberg.listOcrBackends()
+    }
+
+    func testListPostProcessors() throws {
+        // List post-processors
+        let result = try Kreuzberg.listPostProcessors()
+    }
+
+    func testListRenderers() throws {
+        // List renderers
+        let result = try Kreuzberg.listRenderers()
+    }
+
+    func testListValidators() throws {
+        // List validators
+        let result = try Kreuzberg.listValidators()
+    }
+
+}

e2e/swift_e2e/Tests/KreuzbergE2ETests/RendererManagementTests.swift

@@ -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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: renderer_management.
+final class RendererManagementTests: XCTestCase {
+    func testRenderersClear() throws {
+        // Clear all renderers and verify list is empty
+        let result = try Kreuzberg.clearRenderers()
+    }
+
+    func testRenderersList() throws {
+        // List all registered renderers
+        let result = try Kreuzberg.listRenderers()
+    }
+
+}

e2e/swift_e2e/Tests/KreuzbergE2ETests/SmokeTests.swift

@@ -0,0 +1,103 @@
+// 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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: smoke.
+final class SmokeTests: XCTestCase {
+    override class func setUp() {
+        super.setUp()
+        let _testDocs = URL(fileURLWithPath: #filePath)
+            .deletingLastPathComponent() // <Module>Tests/
+            .deletingLastPathComponent() // Tests/
+            .deletingLastPathComponent() // swift/
+            .deletingLastPathComponent() // packages/
+            .deletingLastPathComponent() // <repo root>
+            .appendingPathComponent("test_documents")
+        if FileManager.default.fileExists(atPath: _testDocs.path) {
+            FileManager.default.changeCurrentDirectoryPath(_testDocs.path)
+        }
+    }
+
+    func testOcrImagePng() throws {
+        // OCR: PNG image extraction with OCR enabled. In WASM this exercises the Uint8Array bridge parameter and Promise await in the generated OcrBackend bridge.
+        let result = try Kreuzberg.extractBytes("images/test_hello_world.png", "image/png", "{}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "image/png")
+        XCTAssertGreaterThanOrEqual(result.content.count, 1)
+        XCTAssertTrue(result.content.contains("Hello") || result.content.contains("World") || result.content.contains("hello") || result.content.contains("world"), "expected to contain at least one of the specified values")
+    }
+
+    func testSmokeDocxBasic() async throws {
+        // Smoke test: DOCX with formatted text
+        let result = try await Kreuzberg.extractFile("docx/fake.docx", "application/vnd.openxmlformats-officedocument.wordprocessingml.document", "{}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "application/vnd.openxmlformats-officedocument.wordprocessingml.document")
+        XCTAssertGreaterThanOrEqual(result.content.count, 20)
+        XCTAssertTrue(result.content.contains("Lorem") || result.content.contains("ipsum") || result.content.contains("document") || result.content.contains("text"), "expected to contain at least one of the specified values")
+    }
+
+    func testSmokeHtmlBasic() async throws {
+        // Smoke test: HTML table extraction
+        let result = try await Kreuzberg.extractFile("html/simple_table.html", "text/html", "{}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "text/html")
+        XCTAssertGreaterThanOrEqual(result.content.count, 10)
+        XCTAssertTrue(result.content.contains("Sample Data Table") || result.content.contains("Laptop") || result.content.contains("Electronics") || result.content.contains("Product"), "expected to contain at least one of the specified values")
+    }
+
+    func testSmokeImagePng() async throws {
+        // Smoke test: PNG image (without OCR, metadata only)
+        let result = try await Kreuzberg.extractFile("images/sample.png", nil, "{\"disable_ocr\":true}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "image/png")
+    }
+
+    func testSmokeJsonBasic() async throws {
+        // Smoke test: JSON file extraction
+        let result = try await Kreuzberg.extractFile("json/simple.json", "application/json", "{}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "application/json")
+        XCTAssertGreaterThanOrEqual(result.content.count, 5)
+    }
+
+    func testSmokePdfBasic() async throws {
+        // Smoke test: PDF with simple text extraction
+        let result = try await Kreuzberg.extractFile("pdf/fake_memo.pdf", "application/pdf", "{}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "application/pdf")
+        XCTAssertGreaterThanOrEqual(result.content.count, 50)
+        XCTAssertTrue(result.content.contains("May 5, 2023") || result.content.contains("To Whom it May Concern"), "expected to contain at least one of the specified values")
+    }
+
+    func testSmokeTxtBasic() async throws {
+        // Smoke test: Plain text file
+        let result = try await Kreuzberg.extractFile("text/report.txt", "text/plain", "{}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "text/plain")
+        XCTAssertGreaterThanOrEqual(result.content.count, 5)
+    }
+
+    func testSmokeXlsxBasic() async throws {
+        // Smoke test: XLSX with basic spreadsheet data including tables
+        let result = try await Kreuzberg.extractFile("xlsx/stanley_cups.xlsx", "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet", "{}")
+        XCTAssertEqual(result.mimeType.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines), "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet")
+        XCTAssertGreaterThanOrEqual(result.content.count, 100)
+        XCTAssertTrue(result.content.contains("Team"), "expected to contain: \("Team")")
+        XCTAssertTrue(result.content.contains("Location"), "expected to contain: \("Location")")
+        XCTAssertTrue(result.content.contains("Stanley Cups"), "expected to contain: \("Stanley Cups")")
+        XCTAssertTrue(result.content.contains("Blues"), "expected to contain: \("Blues")")
+        XCTAssertTrue(result.content.contains("Flyers"), "expected to contain: \("Flyers")")
+        XCTAssertTrue(result.content.contains("Maple Leafs"), "expected to contain: \("Maple Leafs")")
+        XCTAssertTrue(result.content.contains("STL"), "expected to contain: \("STL")")
+        XCTAssertTrue(result.content.contains("PHI"), "expected to contain: \("PHI")")
+        XCTAssertTrue(result.content.contains("TOR"), "expected to contain: \("TOR")")
+        // skipped: field 'tables' not available on result type
+        // skipped: field 'metadata.format.excel.sheet_count' not available on result type
+        // skipped: field 'metadata.format.excel.sheet_names' not available on result type
+    }
+
+}

e2e/swift_e2e/Tests/KreuzbergE2ETests/TestHelpers.swift

@@ -0,0 +1,127 @@
+// 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
+// swift-format-ignore-file
+
+import Foundation
+#if canImport(FoundationNetworking)
+// URLSession, URLRequest, HTTPURLResponse, and URLSessionTaskDelegate live in
+// the FoundationNetworking submodule on swift-corelibs-foundation (Linux). On
+// Apple platforms these types remain in plain Foundation and this submodule
+// does not exist; the canImport guard skips the import there.
+import FoundationNetworking
+#endif
+import RustBridge
+
+// Make `RustString` print its content in XCTest failure output. Without this,
+// every error thrown from the swift-bridge layer surfaces as
+// `caught error: "RustBridge.RustString"` with the actual message hidden
+// inside the opaque class instance. The `@retroactive` keyword acknowledges
+// that the conformed-to protocol (`CustomStringConvertible`) and the
+// conforming type (`RustString`) both live outside this module — required by
+// Swift 6 to silence the retroactive-conformance warning. swift-bridge does
+// not give `RustString` a `description` of its own, so there is no conflict.
+extension RustString: @retroactive CustomStringConvertible {
+    public var description: String { self.toString() }
+}
+
+// Spawns the alef mock-server once per test process and exposes its base URL.
+// SwiftPM/XCTest has no global "before all tests" hook that can inject environment
+// variables (the JVM-style listener trick used by the Java/Kotlin backends), so the
+// server is started lazily on first access of `baseURL` and kept alive for the
+// lifetime of the process. A pre-set `MOCK_SERVER_URL` (e.g. exported by CI) wins.
+enum AlefE2EMockServer {
+    static let baseURL: String = AlefE2EMockServer.start()
+
+    // Retain the child process so it is not reaped while tests run.
+    nonisolated(unsafe) private static var process: Process?
+
+    private static func start() -> String {
+        if let preset = ProcessInfo.processInfo.environment["MOCK_SERVER_URL"], !preset.isEmpty {
+            return preset
+        }
+        let fileManager = FileManager.default
+        var dir = URL(fileURLWithPath: fileManager.currentDirectoryPath)
+        var fixturesDir: URL?
+        for _ in 0..<16 {
+            let candidate = dir.appendingPathComponent("fixtures")
+            var isDir: ObjCBool = false
+            if fileManager.fileExists(atPath: candidate.path, isDirectory: &isDir), isDir.boolValue {
+                fixturesDir = candidate
+                break
+            }
+            let parent = dir.deletingLastPathComponent()
+            if parent.path == dir.path { break }
+            dir = parent
+        }
+        guard let fixtures = fixturesDir else {
+            fatalError("AlefE2EMockServer: could not locate fixtures/ above \(fileManager.currentDirectoryPath)")
+        }
+        let repoRoot = fixtures.deletingLastPathComponent()
+        let binary = repoRoot.appendingPathComponent("e2e/rust/target/release/mock-server")
+        guard fileManager.fileExists(atPath: binary.path) else {
+            fatalError("AlefE2EMockServer: mock-server binary not found at \(binary.path) — run: cargo build --manifest-path e2e/rust/Cargo.toml --bin mock-server --release")
+        }
+        let proc = Process()
+        proc.executableURL = binary
+        proc.arguments = [fixtures.path]
+        let stdoutPipe = Pipe()
+        proc.standardOutput = stdoutPipe
+        // Keep stdin open so the server does not see EOF and exit immediately.
+        proc.standardInput = Pipe()
+        do {
+            try proc.run()
+        } catch {
+            fatalError("AlefE2EMockServer: failed to start mock-server: \(error)")
+        }
+        process = proc
+        let handle = stdoutPipe.fileHandleForReading
+        var buffer = Data()
+        var resolved: String?
+        for _ in 0..<500 {
+            let chunk = handle.availableData
+            if chunk.isEmpty { break }
+            buffer.append(chunk)
+            if let text = String(data: buffer, encoding: .utf8) {
+                for line in text.split(separator: "\n") {
+                    if line.hasPrefix("MOCK_SERVER_URL=") {
+                        resolved = String(line.dropFirst("MOCK_SERVER_URL=".count)).trimmingCharacters(in: .whitespacesAndNewlines)
+                        break
+                    }
+                }
+            }
+            if resolved != nil { break }
+        }
+        guard let url = resolved else {
+            proc.terminate()
+            fatalError("AlefE2EMockServer: mock-server did not emit MOCK_SERVER_URL")
+        }
+        // Drain remaining stdout in the background so a full pipe never blocks the server.
+        DispatchQueue.global(qos: .background).async {
+            while !handle.availableData.isEmpty {}
+        }
+        return url
+    }
+}
+
+// URLSession that does not follow redirects, so tests can assert on 3xx status codes
+// and Location headers instead of transparently chasing them to the final response.
+final class AlefE2ENoRedirectDelegate: NSObject, URLSessionTaskDelegate {
+    func urlSession(
+        _ session: URLSession,
+        task: URLSessionTask,
+        willPerformHTTPRedirection response: HTTPURLResponse,
+        newRequest request: URLRequest,
+        completionHandler: @escaping (URLRequest?) -> Void
+    ) {
+        completionHandler(nil)
+    }
+}
+
+let alefE2ESession = URLSession(
+    configuration: .ephemeral,
+    delegate: AlefE2ENoRedirectDelegate(),
+    delegateQueue: nil
+)

e2e/swift_e2e/Tests/KreuzbergE2ETests/ValidatorManagementTests.swift

@@ -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
+// swift-format-ignore-file
+
+import XCTest
+import Foundation
+#if canImport(FoundationNetworking)
+import FoundationNetworking
+#endif
+import Kreuzberg
+import RustBridge
+
+/// E2e tests for category: validator_management.
+final class ValidatorManagementTests: XCTestCase {
+    func testValidatorsClear() throws {
+        // Clear all validators and verify list is empty
+        let result = try Kreuzberg.clearValidators()
+    }
+
+    func testValidatorsList() throws {
+        // List all registered validators
+        let result = try Kreuzberg.listValidators()
+    }
+
+}