# tjmlabs/ColiVara

Colivara is a suite of services that allows you to store, search, and retrieve documents based on their visual embedding. ColiVara has state of the art retrieval performance on both text and visual documents. using vision models instead of chunking and text-processing for documents. No OCR, no text extraction, no broken tables, or missing images.

Repository: https://github.com/tjmlabs/ColiVara
Canonical: https://ross.abutalabs.com/products/colivara
Homepage: https://colivara.com
Language: Python
License: NOASSERTION
License Family: other
Last push: 2026-07-08T15:56:00+00:00

## Health v2 (maintenance only)
Score: 65/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 91, release rhythm 40, longevity 50
- inputs: {"age_days": 712, "days_push": 56, "days_rel": 573, "gap_med": 1, "n_releases_24m": 28}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1486, forks 117 (observed 2026-08-28T04:04:51.932220+00:00)

## What it is
ColiVara is a hosted (and self-hostable) retrieval API that stores, searches, and retrieves documents using visual embeddings generated by vision models like ColQwen2, based on the ColPali paper. It replaces traditional OCR, chunking, and text extraction pipelines, supporting 100+ file formats with Python and TypeScript SDKs.

## Use cases
- build a RAG application that understands tables, charts, and diagrams
- search PDFs and documents without OCR or text extraction
- retrieve documents by visual similarity using embeddings
- index complex financial reports and technical diagrams for semantic search
- filter document search results by metadata and collections
- avoid broken tables and missing images in document retrieval

## When to choose
- your documents are visually rich with tables, figures, or complex layouts that OCR pipelines break
- you want state-of-the-art document retrieval without building a chunking/embedding pipeline yourself
- you need a simple REST API with Python/TypeScript SDKs for document search
- you want to self-host a ColPali-style visual retrieval service

## When to avoid
- you only need plain-text keyword search over simple text documents
- you need full control over chunking and embedding strategies
- you cannot send documents to a third-party API and cannot self-host
- you need a lightweight local library rather than a hosted service

## Facets
- artifact type: service
- maturity: active
- function: rag, vector-database, search-engine, ocr, machine-learning, sdk, api-framework
- domain: large-language-models, artificial-intelligence, pdf, databases
- platform: self-hosted, python
- tags: visual-embeddings, colpali, document-retrieval, vision-models, multimodal-rag, document-search, retrieval-augmented-generation, search, web-server, nodejs, docker

## Member repositories
- tjmlabs/ColiVara (main) score 65

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:51.932220+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:33:53.353065+00:00, confidence not recorded.
  - readme: https://github.com/tjmlabs/ColiVara (fetched 2026-08-28T04:04:51.932220+00:00, sha 0a1159d8cafe)
  - homepage: https://colivara.com (fetched 2026-08-29T11:40:07.310830+00:00, sha 9f29aca94b68)
  - site_page: https://docs.colivara.com (fetched 2026-08-29T11:40:07.314323+00:00, sha e84c3004a399)
- Data as of 2026-08-30T08:39:29.467469+00:00.