We are excited to announce the official launch of chromadb-repo-indexer, released on July 31, 2026. This tool provides a powerful and efficient way to synchronize the text content of your GitHub repositories directly into a remote ChromaDB collection, creating a high-fidelity knowledge base for Retrieval Augmented Generation (RAG) applications.

Key Capabilities

chromadb-repo-indexer is built to handle the complexities of codebase indexing with precision:

  • Context-Aware Chunking: Instead of naive splitting, the indexer uses Markdown-aware and code-aware chunking. By leveraging Tree-sitter, it prefers top-level declarations in code and respects heading structures in Markdown, ensuring that the retrieved context remains meaningful.
  • Namespace Isolation: To support multi-tenant or multi-branch environments, the tool implements a unique namespace identity for every repository and branch combination. This allows you to host content from various sources in a single collection without risk of collision.
  • Safe Synchronization: Data integrity is a priority. The indexer calculates a precise diff between your local files and remote records, performing upserts for new or modified content first, and only deleting stale records after a successful update.
  • Flexible Embeddings: While it supports the ChromaDB server’s native embedding functions, the tool also introduces support for client-side OpenAI-compatible embedding APIs, giving you full control over your vectorization pipeline.

Why It Matters

For developers building LLM-powered tools, the quality of the retrieved context is the primary bottleneck. By automating the synchronization of your repository and using structured chunking, chromadb-repo-indexer ensures that your RAG pipeline always has access to the latest, most relevant version of your documentation and source code.

The integration with GitHub Actions means your vector database stays in sync with your code in real-time, removing the need for manual indexing scripts or complex custom pipelines.

Getting Started

GitHub Action

Integrate the indexer into your CI/CD workflow to automate updates. Note that because the action now runs as a high-performance composite action, you must include a checkout step:

steps:
  - uses: actions/checkout@v4 # Required for v1.0.0+
  - uses: UnitVectorY-Labs/chromadb-repo-indexer@v1
    with:
      server-url: $
      collection-name: repository-content

Python CLI

For local indexing or custom scripts, you can install the CLI via pip:

python -m pip install .

This post was AI-generated using the model unsloth/gemma-4-31B-it-GGUF:UD-Q5_K_XL. Reference: UnitVectorY-Labs/chromadb-repo-indexer, release v1.0.0, generated on August 02, 2026. Author: release-storyteller