CocoIndex

CocoIndex turns codebases, meeting notes, inboxes, Slack, PDFs and videos into continuously updated context for AI agents and LLM apps. You declare in Python what should end up in your target, such as a vector database or Postgres, and CocoIndex keeps it in sync by reprocessing only what changed.

CocoIndex is incremental, so when a source changes only the difference is recomputed, and functions are cached by their input and code. It runs in parallel by default and scales from a laptop to large data. Connectors cover sources such as local files and targets such as Postgres, and there are more than 20 examples to start from.

CocoIndex handles the unglamorous part of RAG and agents: keeping indexes up to date without rebuilding everything. It's Apache-licensed.

Stars

0

Forks

0

Commits

0

Repository

cocoindex-io/cocoindex

• FivetranCommercial managed data pipelines.
• LlamaIndexFramework for connecting LLMs to data, with a hosted platform.
• UnstructuredPlatform for preparing documents for LLMs.