pgvector RAG Backend
by Kitworks Systems
850,00 €
| Odoo Apps Dependencies | base, base_automation, bus, http_routing, mail |
| Community Apps Dependencies | Show |
| Technical Name | kw_ai_rag_pgvector |
| License | OPL-1 |
| Website | https://kitworks.systems/ |
| Versions | 17.0 18.0 19.0 |
AI Connector — RAG Backend
pgvector — PostgreSQL Vector Backend
Run semantic search directly in your existing PostgreSQL database. Zero additional infrastructure — your Odoo server is all you need.
0
Extra Services
HNSW
Index Type
cosine
Similarity Metric
When to Use pgvector
You already run PostgreSQL
Odoo's own database doubles as your vector store. No extra servers, no extra licensing, no extra backups.
Up to ~100 k vectors
HNSW index delivers sub-millisecond recall for most enterprise knowledge bases. Scales without tuning for typical Odoo workloads.
Simplicity is the priority
One technology stack for your entire application. Backup, replication, and monitoring follow existing PostgreSQL procedures.
Technical Specifications
| Property | Value | Notes |
|---|---|---|
| PostgreSQL version | 13+ | pgvector 0.7+ extension required |
| Index type | HNSW | Created automatically on first RAG index run |
| Similarity metric | Cosine | Configurable: L2, inner product |
| Recommended scale | Up to 100 k vectors | Beyond this, consider Qdrant or FAISS |
| External service | None | Uses Odoo's own PostgreSQL |
Requirements
- ✓ kw_ai — AI Connector core
- ✓ kw_ai_rag — RAG orchestration module
- ✓ PostgreSQL 13 or newer with pgvector extension installed
- ✓ Any embedding module (kw_ai_embed_openai, kw_ai_embed_google, or kw_ai_embed_ollama)
Ready to enable semantic search?
No new infrastructure. Works with your existing Odoo PostgreSQL.
Book a Demo