Convert documents to markdown
Drop in Word, PDF, PowerPoint, Excel, and nine more formats. Conversion runs entirely in your browser, so files never leave your machine, and each one takes milliseconds.
How it works
This tool is built on anydoc, an open source Rust library by the Firecrawl team, compiled to WebAssembly so it runs inside your browser instead of on a server. That's where the speed comes from: there's no upload, no queue, and no round trip. The file goes from your disk into memory, through the parser, and out as markdown before a network request would have finished connecting.
Your files stay on your machine. Nothing is sent to us or to anyone else.
Supported formats
Password-protected files and scanned, image-only PDFs are the two things it can't handle. It will say so instead of giving you an empty result.
Why markdown
Markdown is what language models read best. Feeding a raw .docx or PDF to an AI tool wastes context on formatting noise and often loses tables and structure. Converting first gives the model clean headings, lists, and tables it can actually reason about. It's the same reasoning behind our own site servingevery page as markdown to AI agents. If you're pasting documents into Claude or ChatGPT, or building a pipeline that ingests them, this is the step that makes the rest work.
For agents and scripts
The drop zone is for people. If you're an AI agent, a script, or a pipeline, the same converter is a free API with no key and no account. There's also an MCP server, so Claude can convert documents for you after a single setup command.
Convert from any script or pipeline
curl -F file=@report.docx https://dawloom.com/api/convertGive the tool to Claude
claude mcp add --transport http dawloom-tools https://dawloom.com/mcpAfter the MCP setup, asking Claude to "convert this report to markdown" just works. Full usage docs live at dawloom.com/api/convert and the spec at /api/openapi.json. API conversions run in memory and are never stored; for fully local conversion, this page is the tool.
Looking for one format?
Questions we actually get
Short answers, no sales language. If yours isn't here, ask an engineer directly.
Ask us anythingWhere do my files go?
Nowhere. The converter is a WebAssembly module that runs inside your browser tab, so your files never touch a server, ours or anyone else's. You can verify this: load the page, disconnect from the internet, and it keeps converting.
Why does the first conversion take a moment?
Your browser downloads the 2.5 MB converter once, on your first interaction with the drop zone. After that, conversions typically finish in a few milliseconds, and the module stays cached for next time.
Can it convert scanned PDFs?
No. It extracts real text, and a scanned PDF is a stack of images with no text in it. Those need OCR, which this tool doesn't do. It tells you honestly when a PDF has nothing to extract instead of returning an empty file.
Is it really free? What is the catch?
It's free and there's no account, no file limit, and no watermark. We're a software agency, and tools like this are how we show our work. If you ever need something bigger built, you know where we are.
Can my AI agent or script use this?
Yes. The same converter runs as a free API at dawloom.com/api/convert (one curl, no key) and as an MCP server at dawloom.com/mcp that you can add to Claude Code or Claude Desktop with one command. Both are documented in the "For agents" section on this page.
Need more than a converter?
We build document pipelines, AI integrations, and full products. An engineer replies, not a sales team.