odoo-ai-agent logo

odoo-ai-agent

Free

Agent IA qui répond à des questions métier sur une base Odoo, avec garde-fous SQL et évaluation de l'ancrage des réponses

Visit Website
Related by shared tags and keywordsotto_agentqwenCerveau
ai-agenterplanggraphllmodooaiai assistantautomation

Overview

odoo-ai-agent is an AI agent that answers business questions about an Odoo database in French, without inventing figures. It addresses three real problems: the model's ignorance of Odoo management rules, the risk of destructive queries, and the possibility of fabricated numbers. The agent uses a BM25-based functional knowledge base to inform the model about Odoo rules, three layers of safeguards including a SQLite authorizer to prevent destructive queries, and an evaluation harness that verifies accuracy and grounding on actual queries. The architecture is a two-node LangGraph with conditional edges, where the 'think' node uses an LLM with tools and the 'act' node executes tools like list_tables, describe_table, run_sql, and search_docs. The agent can be used via CLI commands such as 'python -m odoo_ai_agent ask' or an interactive shell, and it supports both SQLite and PostgreSQL databases with read-only access. The project includes 62 tests with no network calls, ensuring deterministic CI.

Pricing

Free

Detailed plans have not been confirmed in our catalog. Check the official website for current limits and billing terms.

Visit Website

Prices and limits may change. Confirm the currency, billing period, seat minimum and usage caps on the official website.

Use Cases

  • Teams use odoo-ai-agent to ask business questions in French about an Odoo database, such as revenue or customer outstanding amounts.
  • Teams use odoo-ai-agent to run offline demonstrations without an API key.
  • Teams use odoo-ai-agent to query a real Odoo database via PostgreSQL with read-only access.
  • Teams use odoo-ai-agent to evaluate the agent's accuracy, grounding, and refusal behavior on a set of questions.
  • Teams use odoo-ai-agent to test the orchestration of the agent with a FakeLLM, ensuring deterministic CI.

Features

AI assistant

Natural language interface

Workflow automation

Reviews

No reviews yet. Be the first to share your experience!