Skip to content

feat: Embedding semantic search with pgvector#154

Open
rivetphilbot wants to merge 4 commits intoMartian-Engineering:mainfrom
rivetphilbot:feat/embedding-semantic-search-v3
Open

feat: Embedding semantic search with pgvector#154
rivetphilbot wants to merge 4 commits intoMartian-Engineering:mainfrom
rivetphilbot:feat/embedding-semantic-search-v3

Conversation

@rivetphilbot
Copy link
Contributor

Replaces #141 (rebased onto postgres-backend-v3 + upstream a2fa42f).

Adds semantic search capabilities:

  • pgvector embedding columns on messages and summaries
  • Background embedding queue with OpenAI API
  • Agent registry with instance tracking
  • Recency-boosted semantic search (0.7 semantic + 0.3 time decay)
  • Backfill scripts for existing data
  • 412 tests passing

Depends on postgres-backend-v3 PR.

Includes upstream fixes:
- SQLite busy_timeout for concurrent access
- CJK LIKE fallback for FTS5
- Summarizer LLM timeout protection
- Sync refCount in withSessionQueue
- Media message annotation in compaction
- Persist summary model in database

Adds model column to postgres schema (CREATE + migration).
…y boost

Rebased onto postgres-backend-v3 (upstream a2fa42f).
Includes all 6 upstream fixes + model column support.
@rivetphilbot rivetphilbot force-pushed the feat/embedding-semantic-search-v3 branch from a33f946 to 5f98e50 Compare March 22, 2026 12:54
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant