10 open-source repos that turn any document into agent-readable text, from MarkItDown to olmOCR
victor_explore · x · 2026-10-17
A curated list of 10 open-source GitHub repos that parse documents into text your AI agent can read (star counts checked Oct 2026):
- MarkItDown (Microsoft, 189K stars, MIT): one command converts Word, PowerPoint, Excel, PDF and HTML to Markdown; image reading needs a plugin and model
- PaddleOCR (91K stars, Apache 2.0): reads scans/photos in 100+ languages, outputs Markdown or JSON, includes a small vision model for hard pages
- MinerU (81K stars, custom license): strong on tables, formulas and Asian languages; paid license for large companies
- Docling (started at IBM, 69K stars, MIT): reads PDF, Office, HTML, email and audio, runs on CPU, ships an MCP server
- Marker (40K stars, Apache 2.0): fast PDF-to-Markdown with math support
The accompanying graphic also covers OpenDataLoader PDF (layout-aware parsing), olmOCR (Allen AI), Unstructured (RAG chunking), LiteParse (lightweight, no GPU) and Xberg (Rust engine).
More from coding & agent
- Superpowers beats Agent Skills in head-to-head test: same app in 49 min for $4.76 vs 90 min for $6.30 — chessbuzz · 2026-10-17
- Pydantic Monty ships TypeScript bindings to sandbox model-written Python in Node and browsers — samuelcolvin · 2026-10-17
- LangChain's deepagents adds Gemini support with 1M tokens and agentic video understanding — hwchase17 · 2026-10-17
- Dev swaps codex remote and Claude remote control for T3 Code mobile — rudrank · 2026-10-17
- Perplexity Flash matches Wald's accuracy at nearly 2x speed in local agent-decision benchmark — WolframRvnwlf · 2026-10-17
- EpochGo open-sources local-first AutoML with offline Ollama LLMs and deterministic pipelines — Grouchy-Excitement28 · 2026-10-17