tjmlabs/ColiVara
A document retrieval service that uses vision models to create visual embeddings for indexing and searching across 100+ file formats including PDFs, DOCX, and PPTX.
Not currently ranked — collecting fresh signals.
star history
ColiVara indexes documents by generating visual embeddings using vision models rather than traditional text chunking or OCR. This approach handles complex document layouts without broken tables or missing images. It provides Python and TypeScript SDKs for document upload, semantic retrieval, and collection management, targeting state-of-the-art retrieval performance for both text-heavy and visual documents.
Frequently asked
- What is tjmlabs/ColiVara?
- A document retrieval service that uses vision models to create visual embeddings for indexing and searching across 100+ file formats including PDFs, DOCX, and PPTX.
- Is ColiVara open source?
- Yes — tjmlabs/ColiVara is an open-source project tracked on heatdrop.
- What language is ColiVara written in?
- tjmlabs/ColiVara is primarily written in Python.
- How popular is ColiVara?
- tjmlabs/ColiVara has 1.5k stars on GitHub.
- Where can I find ColiVara?
- tjmlabs/ColiVara is on GitHub at https://github.com/tjmlabs/ColiVara.