Algis DumbrisandClaude Opus 4.6 5dcb9e6149 feat: implement semantic search with embedding pipeline ...

Add HNSW-based vector search with configurable embedding providers
(OpenAI, Ollama) and automatic FTS5 fallback when no provider is
configured. Background pipeline embeds messages asynchronously on
ingest, stores vectors in a pure-Go HNSW index, and retries on
failure with exponential backoff. The search_messages MCP tool now
supports search_mode (auto/semantic/fulltext) and returns ranked
results with similarity scores. All existing tests continue to pass,
CGO_ENABLED=0 cross-compilation verified.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

2026-03-13 12:14:36 +02:00

..

001_initial.sql

feat: implement core foundation (storage, messaging, agents, MCP server)

2026-03-13 11:45:22 +02:00

002_channels.sql

feat: implement channels with membership and MCP tools

2026-03-13 11:56:50 +02:00

003_trace_logging.sql

fix: resolve migration version conflict (channels=002, trace=003)

2026-03-13 12:01:26 +02:00

004_auth.sql

feat: merge auth, channels, and trace logging (Round 2)

2026-03-13 12:02:47 +02:00

005_semantic_search.sql

feat: implement semantic search with embedding pipeline

2026-03-13 12:14:36 +02:00