Aura GraphRAG Backend
by Kitworks Systems
650,00 €
| Odoo Apps Dependencies | base, base_automation, bus, http_routing, mail |
| Community Apps Dependencies | Show |
| Technical Name | kw_ai_aura |
| License | OPL-1 |
| Website | https://kitworks.systems/ |
| Versions | 17.0 18.0 19.0 |
AI Connector — RAG Backend
Aura — Vector + Graph Memory Backend
Graph-augmented retrieval for AI agents that need to reason across relationships, not just retrieve similar text. Combines vector similarity with knowledge graph traversal.
vector
+ graph
GraphRAG
Architecture
agent
Optimised For
When to Use Aura
Relationship-aware retrieval
When answers depend on traversing connections — partner → orders → products → suppliers — not just matching text similarity.
Long-term agent memory
Agents store and recall facts as graph nodes. Context persists across sessions without growing the prompt window.
Multi-hop reasoning
Find answers that require combining information from multiple Odoo records via explicit relationship edges.
Technical Specifications
| Property | Value | Notes |
|---|---|---|
| Architecture | GraphRAG (vector + graph) | Powered by AuraSDK Python library |
| Memory scope | Per-agent, per-user | Isolated namespaces |
| Retrieval strategy | Similarity + graph traversal | Configurable hop depth |
| External dependency | aura Python library | Not in kw_ai_bundle due to this dependency |
| Agent integration | kw_ai_agent + kw_ai_memory | Replaces default memory backend |
Requirements
- ✓ kw_ai — AI Connector core
- ✓ kw_ai_rag — RAG orchestration module
- ✓ kw_ai_agent + kw_ai_memory — for agent-scoped graph memory
- ✓
auraPython library installed in Odoo virtualenv - ✓ Any embedding module
Agents that understand relationships, not just text
GraphRAG for Odoo — when vector similarity alone is not enough.
Book a Demo