
mdstill
Clean Markdown for ChatGPT, Claude, Gemini and RAG
Details
- Categories
- AIDeveloper ToolsData & Infrastructure
- Target Audience
- AI DevelopersSoftware DevelopersNon-Technical Users
- Pricing
- Free
- Alternative To
Unstructured.io
Firecrawl
About mdstill
mdstill is a document preprocessor purpose-built for LLM and RAG workflows. Drop a PDF, Word, Excel, PowerPoint, EPUB, or any of 18 supported formats and get back clean, structure-preserving Markdown that ChatGPT, Claude, Gemini, and vector-database pipelines can actually use. Key features: • 18 input formats: PDF, DOCX, PPTX, XLSX, EPUB, HTML, CSV, JSON, RTF, ODT, Apple iWork (Pages, Numbers, Keynote) and more • GitHub-flavored Markdown output — tables preserved, headings become linkable anchors, lists and code blocks survive intact • ~5-30% more token-efficient than raw text extraction — your LLM costs drop accordingly • REST API for pipeline automation and batch workflows • Free tier with no signup for basic use; registered users get higher daily limits • Privacy-first: files processed in-memory and deleted immediately after conversion — no storage, no logging, no training Built for engineers shipping AI features (RAG ingestion, agent context, embedding pipelines) and for knowledge workers importing legacy archives into Obsidian, Notion, or Logseq. The same engine powers both the zero-install web tool and the developer API. Alternative to: markitdown, Unstructured.io, LlamaParse, Docling, Firecrawl.
Product Insights
This web and API-based developer tool converts 18 file formats into structured, token-efficient Markdown optimized for LLM ingestion and RAG pipelines. It provides a privacy-first, free-to-use utility that preserves document structure for AI agents and knowledge management systems.
- Supports 18 formats including PDF, Excel, and Apple iWork documents.
- Improves token efficiency by 5-30% compared to raw text extraction.
- Ensures data privacy by processing files in-memory without persistent storage.
- Offers a REST API for automated batch workflows and RAG ingestion.
Ideal for: AI Developers and Software Developers who need to prepare structured document data for RAG pipelines, AI agents, and knowledge bases.
This tool serves as an alternative to markitdown, Unstructured.io, LlamaParse, Docling, and Firecrawl.
Screenshots
Reviews (0)
No reviews yet. Be the first to rate this product!










Comments (1)
No single tool handled all my document formats — PDF, Word, Excel, EPUB, and more — when I needed to feed them into an LLM. So I built mdstill. 18 formats in, clean Markdown out. Free, no signup.