PDF, Word and Excel to clean Markdown in one API call
Send up to 5 documents per call, by public link or as an upload (10 MB each), and get back Markdown with headings and tables kept, plus document metadata, character and approximate token counts, a one-line preview and a flag when a PDF has no text layer. It works as a REST API with an OpenAPI file and as a remote MCP server, so AI agents, RAG pipelines and no-code tools can call it without installing a Python library. Documents are converted in memory and never stored.
Live calls to the Leafmelt API, 27 Sep 2026: 3 sample files plus 2 public documents
| Document | Format | Size | Characters | ≈ Tokens | Time | Note |
|---|---|---|---|---|---|---|
| quarterly-report.pdf | 2.5 KB | 520 | 130 | 0.5 s | table + headings | |
| meeting-notes.docx | Word | 37 KB | 467 | 117 | 0.6 s | Markdown table kept |
| orders.xlsx | Excel | 5.5 KB | 430 | 108 | 0.5 s | one table per sheet |
| arXiv 1706.03762 (15 pages) | 2.2 MB | 39,413 | 9,854 | 1.8 s | full paper, not cut | |
| IRS Form W-9 (XFA form) | 141 KB | 85 | 22 | 0.8 s | emptyText: no text layer |
Source: products/leafmelt/store/listing.json (sample table from live API calls on 27 Sep 2026; the three sample files are fictional and served at leafmelt.cybermaxtools.com/samples/).
Leafmelt vs the alternatives
Published prices, each checked on the date shown; prices change, so confirm on each site.
| Product | Price | Free | Checked |
|---|---|---|---|
| Leafmelt API | $5 once for 750 calls; $19/month for 3,000 calls (up to 5 documents per call) | 50 calls a day, no key | live |
| LlamaParse (LlamaCloud) | Starter $50/month (40K credits), Pro $500/month; 1,000 credits = $1.25, basic parsing from 1 credit per page | 10K credits a month | 2026-10-10 |
| Unstructured | $0.015 per page pay as you go; Business custom | first 10,000 pages, no card | 2026-10-10 |
| Reducto | Parse $10 per 1,000 pages (Standard pay as you go) | $150 of free usage on Standard | 2026-10-10 |
| Mathpix Convert API | PDF $0.005 per page plus a $19.99 one-time setup fee | $29 test credit | 2026-10-10 |
Why teams pick Leafmelt
- Twelve input formats in one endpoint: PDF, Word (.docx), four Excel variants, OpenDocument text and sheets, Apple Numbers, CSV, HTML and XML.
- Word and Excel tables come back as Markdown tables, one table per sheet, so an LLM can read rows and columns.
- A token count on every document, so you can budget context before you send it to a model.
- A daily free tier (50 calls) and a $5 pack that never expires; no subscription needed to start.
- REST, OpenAPI and a remote MCP server on the same key: agents call it as a tool, scripts call it as a URL.
Try it in one call
curl 'https://leafmelt.cybermaxtools.com/api/convert?url=https://leafmelt.cybermaxtools.com/samples/quarterly-report.pdf'How it works
- Open the Try-it box and convert one of the sample files, or paste a link to your own document.
- Call /api/convert with up to 5 document URLs (or POST base64 files); 50 calls a day need no key.
- When you need more, buy the $5 pack or a plan; the key is shown right after checkout.
FAQ
Is there a free PDF to Markdown API?
Yes. Leafmelt gives 50 calls a day per IP with no key and no card, and each call can convert up to 5 documents.
Why not LlamaParse?
Pick LlamaParse when you need OCR or layout models for scanned pages and complex layouts. For documents that already have a text layer, Leafmelt is cheaper to start ($5 once, or $19/month) and also takes Word, Excel and OpenDocument files.
Why not markitdown or docling?
They are good free Python libraries, but they need Python and a server. Leafmelt is one HTTPS call or an MCP tool, so remote agents and no-code tools can use it.
Does it do OCR?
No. Images and scanned PDFs without a text layer are not converted; the response says so with emptyText: true. Old Word .doc files must be saved as .docx first, and lines inside a PDF paragraph can run together.
Are my documents stored?
No. Documents are converted in memory and never stored.