| Availability |
Odoo Online
Odoo.sh
On Premise
|
| Community Apps Dependencies | Show |
| Lines of code | 6958 |
| Technical Name |
angkor_ai_agent |
| License | LGPL-3 |
Angkor AI Agent
AI chat assistant for Odoo that helps users query records, inspect model fields, generate summaries, list PDF reports, download reports, and prepare API Gateway roles.
What It Does
Chat With Odoo Data
Ask for customers, users, employees, invoices, products, sales orders, projects, and other Odoo records.
Render Results Clearly
View responses as user-friendly tables or switch to raw JSON for technical inspection.
Download PDF Reports
List QWeb PDF reports and render report files such as employee badge PDFs with download links.
Key Features
- Backend Odoo chat page with light and dark themes.
- OpenAI-compatible provider support.
- Ollama local model support.
- Fallback deterministic planner when no LLM is available.
- Odoo ORM query execution with validated domains and fields.
- Grouped reports through Odoo read_group.
- Model field inspection for safer query building.
- PDF report listing and download URL generation.
- API role generation for Angkor API Gateway.
- Separate AI Agent User group for normal users.
How Users Work
Example Messages
show latest customers
inspect employee fields
get all invoice that have due today
get vendor bills due tomorrow
report sale order total by state
get list all report for employee
download badge report for employee where ids = 2
create read api role for res.partner
Context JSON Examples
{
"model": "res.partner",
"fields": ["id", "name", "email"],
"limit": 20,
"order": "id desc"
}
{
"ids": [2],
"filename": "Badge - aa.pdf"
}
REST API
The same agent can be called through Angkor API Gateway with a bearer token.
curl -X POST http://localhost:8069/api_gateway/v1/ai_agent/chat \
-H "Authorization: Bearer <access_token>" \
-H "Content-Type: application/json" \
-d '{"message":"Show latest customers","context":{"model":"res.partner","fields":["name","email"],"limit":10}}'
Providers
Ollama Local
Use a local Ollama server such as http://localhost:11434 with a model like llama3.1.
OpenAI Compatible
Use OpenAI or another compatible chat-completions API with your configured API key.
No LLM
Run deterministic fallback planning for supported patterns without an external model.
Please log in to comment on this module