Looking for a fast and frictionless way to turn standard documents into beautifully formatted Markdown? Look no further! MarkItDown is a versatile AutoHotkey v2 script that lets you drag and drop files like PDFs, Word docs, Excel sheets, and even images to quickly generate clean Markdown text. Best of all, it supports optional AI integrations so you can effortlessly extract text from images and transcribe audio. Plus, by converting documents locally first, the tool maximizes efficiency—complete with a visual bar chart that shows exactly how many AI tokens you’re saving! Download the tool today to completely streamline your note-taking, documentation, and content extraction workflow!
🚀 This PDF to Markdown Tool Saves 94% of AI Tokens (Game Changer!) |AHK Hero Extract
Turn Any File Into Clean Markdown — In Seconds
MarkItDown is a drag-and-drop Windows tool that converts PDFs, Word docs, spreadsheets, webpages, emails, images, audio, and more into clean Markdown that is ready to paste into Claude, ChatGPT, Gemini, Obsidian, or any vector store. No command line. No cloud uploads (unless you want them). Just drop and go.
Why Feed AI a 20 MB PDF When You Can Feed It Markdown?
Binary documents are bloated, noisy, and expensive to send through an LLM API. A PDF packed with headers, footers, fonts, and base64 images eats thousands of unnecessary tokens before the model sees a single useful word. Markdown is the opposite: clean text, preserved structure, a fraction of the size.
MarkItDown converts in one click and even shows you — with a bar chart — how many tokens you just saved.
What You Get
- Drag-and-Drop Simplicity — Drop a file, a dozen files, or an entire folder. Subfolders scan automatically.
- 20+ Formats Supported Locally — PDF, DOCX, PPTX, XLSX, XLS, HTML, CSV, JSON, XML, EPUB, MSG (Outlook), IPYNB (Jupyter), ZIP, TXT, MD, RST, LOG, RSS, ATOM — all handled on your machine.
- Optional AI Vision & Audio — Add an OpenRouter key to caption images (JPG, PNG, GIF, BMP, TIFF) and transcribe audio (MP3, WAV, M4A) using GPT-4o, Claude, or Gemini.
- Token Savings Chart — After every batch, see exactly how many API tokens you just saved. Typical savings: 60–90%.
- Bring Your Own Model — Pick GPT-4o, GPT-4o-mini, Claude Sonnet, Claude Haiku, or Gemini 2.5 Flash from a dropdown.
- Customizable LLM Prompt — Tune exactly how images get described. Ships with a detailed default that preserves tables, headings, and lists.
- Smart Batch Retry — Failed files stay in the list so you can reprocess just those with one click.
- Central Output Folder — Point every conversion at one folder to build a searchable Markdown library of your whole document collection.
- 100% Offline for Documents — Nothing is uploaded for PDF/DOCX/etc. conversions. AI is opt-in, per run, via a checkbox.
- Portable — Settings live in a single
markitdown.ininext to the exe. Copy the folder to a USB stick and go.
Every Format It Handles
Documents, web pages, data, books, email — all local, no account needed. Media files unlock with an optional OpenRouter key.
| Category | Formats | AI Key? |
|---|---|---|
| Documents | PDF, DOCX, PPTX, XLSX, XLS | — |
| Web & Data | HTML, HTM, CSV, JSON, XML, RSS, ATOM | — |
| Books & Email | EPUB, MSG (Outlook), IPYNB, ZIP | — |
| Plain Text | TXT, MD, RST, LOG | — |
| Images | JPG, JPEG, PNG, GIF, BMP, TIFF | ✓ |
| Audio | MP3, WAV, M4A | ✓ |
Who It’s For
- AI Power Users — Feed LLMs clean text, cut your token bill, and get better answers from the same models.
- Researchers & Students — Convert reading lists of PDFs and EPUBs into searchable Markdown notes.
- Knowledge Workers — Turn spreadsheets, email threads, and PowerPoints into text you can actually paste into a chat.
- Obsidian / Notion Users — Bulk-ingest old documents into your knowledge system.
- Developers — Preprocess documents into Markdown before chunking and embedding for RAG.
How It Works
- Launch MarkItDown. The main window opens with a file list and three buttons.
- Drop files or folders onto the window (or use
Ctrl+Oto pick files,Ctrl+Shift+Ofor a whole folder). - Click Convert All. Watch the progress bar — the Status column updates per file.
- See your token savings in the bar-chart popup that appears after the batch.
- Click Open Output to see the Markdown files.
What Makes It Different
- You own your data. Documents stay on your computer. AI is opt-in per run.
- You own your model. Plug in any OpenRouter-compatible model — no vendor lock-in.
- You see the savings. Most tools hide token usage. MarkItDown charts it for you.
- Real batch support. Folder recursion, duplicate detection, failure retry — built for volume.
MarkItDown is built by the-Automator.com. Requires Windows 10 or 11. Version 0.1.0.
