Vibe Coding Discoverchunky
View on GitHubOpen-source toolkit for reliable RAG pipelines: convert PDFs to Markdown, clean documents, inspect chunks, compare chunking strategies, and enrich metadata for LLM applications.
A local workspace for preparing documents for RAG. Converts PDFs to Markdown, cleans and reviews documents, compares chunking strategies, saves chunk versions, and uses LLMs to enrich Markdown and chunks.
Use Cases
Convert PDFs into Markdown for RAG pipelinesCompare chunking strategies on source documentsInspect and validate chunks before indexingClean OCR and PDF conversion artifactsEnrich chunks with titles, summaries, keywords, and retrieval questionsProcess and enrich batches of documents
Built With
- Language
- Python
- Frameworks
- Chonkie · LangChain · Docling · FastAPI · Ollama · OpenAI-compatible API
Tags
RAG · document processing · PDF to Markdown · chunking · chunk inspection · chunk evaluation · metadata enrichment · Markdown cleanup · semantic chunking · local LLM