@wsqc2026/markitdown-typescript 0.1.0

package/dist/markitdown.d.ts

@@ -0,0 +1,16 @@
+import type { Converter } from "./converter.js";
+import type { ConvertOptions, ConvertResult, StreamInfo } from "./types.js";
+export interface MarkItDownOptions extends ConvertOptions {
+}
+export interface ConvertInput {
+    streamInfo?: StreamInfo;
+}
+export declare function createMarkItDown(options?: MarkItDownOptions): {
+    convert: (source: string | Buffer, input?: ConvertInput & {
+        _zipDepth?: number;
+    }) => Promise<ConvertResult>;
+    convertLocal: (filePath: string, streamInfo?: StreamInfo) => Promise<ConvertResult>;
+    convertUri: (uri: string, streamInfo?: StreamInfo) => Promise<ConvertResult>;
+    convertUrl: (url: string, streamInfo?: StreamInfo) => Promise<ConvertResult>;
+    registerConverter: (conv: Converter, priority?: number) => void;
+};

package/dist/stream-info.d.ts

@@ -0,0 +1,14 @@
+import type { StreamInfo } from "./types.js";
+/**
+ * Merge multiple StreamInfo objects. Later values override earlier ones,
+ * but only for non-undefined fields.
+ */
+export declare function mergeStreamInfo(base: StreamInfo, ...overrides: (StreamInfo | undefined)[]): StreamInfo;
+/**
+ * Guess mimetype from extension using a built-in lookup table.
+ */
+export declare function guessMimeFromExtension(ext: string): string | undefined;
+/**
+ * Guess extension from mimetype.
+ */
+export declare function guessExtensionFromMime(mime: string): string | undefined;

package/dist/transforms/decode-text.d.ts

@@ -0,0 +1,6 @@
+/**
+ * Decode a buffer to string using the given charset.
+ * Falls back to utf-8 if no charset is specified.
+ * Uses iconv-lite for non-standard encodings.
+ */
+export declare function decodeBuffer(buffer: Buffer, charset?: string): string;

package/dist/transforms/html-to-markdown.d.ts

@@ -0,0 +1,5 @@
+import type { ConvertOptions } from "../types.js";
+export declare function htmlToMarkdown(html: string, opts?: ConvertOptions): {
+    markdown: string;
+    title?: string;
+};

package/dist/types.d.ts

@@ -0,0 +1,26 @@
+import { z } from "zod/v4";
+export declare const StreamInfoSchema: z.ZodObject<{
+    mimetype: z.ZodOptional<z.ZodString>;
+    extension: z.ZodOptional<z.ZodString>;
+    charset: z.ZodOptional<z.ZodString>;
+    filename: z.ZodOptional<z.ZodString>;
+    localPath: z.ZodOptional<z.ZodString>;
+    url: z.ZodOptional<z.ZodString>;
+}, z.core.$strip>;
+export type StreamInfo = z.infer<typeof StreamInfoSchema>;
+export declare const ConvertResultSchema: z.ZodObject<{
+    markdown: z.ZodString;
+    title: z.ZodOptional<z.ZodString>;
+}, z.core.$strip>;
+export type ConvertResult = z.infer<typeof ConvertResultSchema>;
+export declare const ConvertOptionsSchema: z.ZodObject<{
+    llmClient: z.ZodOptional<z.ZodAny>;
+    llmModel: z.ZodOptional<z.ZodString>;
+    llmPrompt: z.ZodOptional<z.ZodString>;
+    exiftoolPath: z.ZodOptional<z.ZodString>;
+    styleMap: z.ZodOptional<z.ZodString>;
+    keepDataUris: z.ZodOptional<z.ZodBoolean>;
+    youtubeTranscriptLanguages: z.ZodOptional<z.ZodArray<z.ZodString>>;
+    _zipDepth: z.ZodOptional<z.ZodNumber>;
+}, z.core.$strip>;
+export type ConvertOptions = z.infer<typeof ConvertOptionsSchema>;

package/dist/uri-utils.d.ts

@@ -0,0 +1,9 @@
+export declare function fileUriToPath(fileUri: string): {
+    netloc: string | null;
+    path: string;
+};
+export declare function parseDataUri(uri: string): {
+    mimeType: string | null;
+    attributes: Record<string, string>;
+    data: Buffer;
+};

package/package.json

@@ -0,0 +1,49 @@
+{
+  "name": "@wsqc2026/markitdown-typescript",
+  "version": "0.1.0",
+  "description": "A TypeScript port of Microsoft's markitdown Python library — converts documents to Markdown",
+  "type": "module",
+  "main": "dist/index.js",
+  "types": "dist/index.d.ts",
+  "bin": {
+    "markitdown": "dist/cli.js"
+  },
+  "files": [
+    "dist",
+    "LICENSE"
+  ],
+  "scripts": {
+    "build": "bun build src/index.ts src/cli.ts --outdir dist --target node --packages external && tsc --emitDeclarationOnly --outDir dist",
+    "compile": "bun build src/cli.ts --compile --outfile markitdown",
+    "test": "bun test",
+    "test:coverage": "bun test --coverage --coverage-reporter=lcov --coverage-reporter=text",
+    "lint": "biome check src/ tests/",
+    "lint:fix": "biome check --write src/ tests/",
+    "typecheck": "tsc --noEmit",
+    "publish:dry": "./scripts/publish.sh --dry-run",
+    "publish:npm": "./scripts/publish.sh"
+  },
+  "dependencies": {
+    "cfb": "^1.2.2",
+    "cheerio": "^1.2.0",
+    "fast-xml-parser": "^5.4.2",
+    "file-type": "^19.6.0",
+    "iconv-lite": "^0.6.3",
+    "jszip": "^3.10.1",
+    "mammoth": "^1.11.0",
+    "pdf-parse": "^2.4.5",
+    "pdfplumber-wasm": "^0.2.0",
+    "turndown": "^7.2.2",
+    "turndown-plugin-gfm": "^1.0.2",
+    "xlsx": "^0.18.5",
+    "youtube-transcript": "^1.2.1",
+    "zod": "^4.0.0-beta.20250505"
+  },
+  "devDependencies": {
+    "@biomejs/biome": "^2.4.10",
+    "@types/bun": "latest",
+    "@types/turndown": "^5.0.6",
+    "typescript": "^5.7.0"
+  },
+  "license": "Apache-2.0"
+}