aim-github-search

Semantic search for GitHub issues and pull requests uses jina-embeddings-v2-base-en vectors stored in a Qdrant database to match content by meaning instead of exact keywords, helping you find conceptually related results.

Yes, you can filter GitHub semantic search results by type and state. Specific payload fields allow you to narrow results to issues, pull requests, commits, or CI results, and filter by open, closed, or merged states.

The best way to find merged pull requests related to authentication bugs is using semantic search with state filtering. Search for the meaning of "authentication bugs" and filter payload fields to pull requests with a merged state.

Yes, semantic GitHub search works with commits and CI results. The system indexes issues, pull requests, commits, and CI results in a Qdrant vector database, allowing you to search and filter across all these content types using vector similarity.

Yes, you need a Qdrant vector database to perform semantic search on GitHub content. The system stores issues, pull requests, commits, and CI results in Qdrant and uses jina-embeddings-v2-base-en to generate vectors for similarity matching.

Limitations of semantic search for GitHub commits and issues include dependency on Qdrant payload fields for type and state filtering, and reliance on jina-embeddings-v2-base-en vector similarity, which may miss exact keyword matches or content not indexed in the database.