Settings UI

The module adds a Gemini AI Connector block in Odoo Settings with:

• Gemini API Key.
• Gemini model selector.
• Custom Gemini model ID.
• Gemini API Base URL.
• Request Timeout.
• Test Gemini Connection action.
• Total Tokens Used counter.
• Monthly Token Limit.
• Reset Counter action.

Reusable Service

The module registers this Odoo abstract model:

sm.gemini.connector.service

Dependent modules can call:

env['sm.gemini.connector.service'].call_ai_chat(messages)

or directly:

env['sm.gemini.connector.service'].call_gemini_chat(messages)

JSON Helper

The connector also includes:

env['sm.gemini.connector.service'].parse_json_response(response)

This helper removes common Markdown code fences and extracts JSON blocks from Gemini responses.

Configuration Parameters

The connector stores settings in ir.config_parameter:

• sm_gemini_connector.gemini_api_key
• sm_gemini_connector.gemini_model
• sm_gemini_connector.gemini_custom_model
• sm_gemini_connector.gemini_api_base_url
• sm_gemini_connector.request_timeout
• sm_gemini_connector.total_tokens_used
• sm_gemini_connector.monthly_token_limit

Dependencies

The module depends on:

• base
• web

No Sales, CRM, Inventory, Accounting, or Helpdesk dependency is required. This is intentional so the connector can be reused by many app types.

Get API Key

1. Open Google AI Studio.
2. Create or select a project.
3. Generate a Gemini API key.
4. Copy the API key.

Configure Odoo

1. Open Settings -> Gemini AI Connector.

2. Enter Gemini API Key.

3. Select a Gemini model.

4. Keep API Base URL as:

   https://generativelanguage.googleapis.com/v1beta
   

5. Set request timeout if needed.

6. Click Save.

7. Click Test Gemini Connection.

Model Selection

The connector currently includes:

• Gemini 2.5 Flash Lite.
• Gemini 2.5 Flash.
• Gemini 2.5 Pro.
• Gemini Flash Latest.
• Gemini 2 Flash.
• Gemini 2 Flash Lite.
• Custom Model ID.

Recommended starting model:

gemini-2.5-flash-lite

This model is usually a practical first choice for demos and lightweight AI automations.

Custom Model ID

Use Custom Model ID when Google exposes a model not listed in the dropdown.

Example:

gemini-2.5-flash-lite

Common Failure Reasons

Basic Chat Call

Use chat-style message dictionaries. The connector converts them to Gemini request format internally.

messages = [
    {
        'role': 'user',
        'content': 'Explain Odoo Sales in one short paragraph.',
    }
]

response = env['sm.gemini.connector.service'].call_ai_chat(
    messages,
    temperature=0.7,
    max_tokens=300,
)

Direct Gemini Call

response = env['sm.gemini.connector.service'].call_gemini_chat(
    messages,
    temperature=0.4,
    max_tokens=500,
)

System Prompt

The connector supports system messages.

messages = [
    {
        'role': 'system',
        'content': 'You are a concise Odoo assistant.',
    },
    {
        'role': 'user',
        'content': 'Summarize this customer note.',
    },
]

response = env['sm.gemini.connector.service'].call_ai_chat(messages)

Model Override

You can override the configured model for one call:

response = env['sm.gemini.connector.service'].call_ai_chat(
    messages,
    model='gemini-2.5-flash',
)

JSON Response Parsing

raw_response = env['sm.gemini.connector.service'].call_ai_chat(messages)
data = env['sm.gemini.connector.service'].parse_json_response(raw_response)

The parser handles:

• Plain JSON.
• Markdown json code fences.
• Markdown generic code fences.
• Text responses that contain a JSON object.

Endpoint

The connector calls:

POST {Gemini API Base URL}/models/{model}:generateContent

Default endpoint:

https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash-lite:generateContent

Headers

The request uses:

Content-Type: application/json
x-goog-api-key: <Gemini API Key>

Payload

The connector converts Odoo/chat-style messages to Gemini payload:

• system -> systemInstruction
• user -> Gemini user role
• assistant -> Gemini model role

Generation Config

The connector sends:

• temperature
• maxOutputTokens
• topP
• topK

Monthly Token Limit

The monthly limit is stored in:

sm_gemini_connector.monthly_token_limit

If the limit is 0, the connector treats it as unlimited.

If total usage reaches the configured limit, the connector stops requests and raises:

Gemini Connector monthly token limit reached.

Reset Counter

Use Reset Counter in Settings to reset usage to 0.

API Key Storage

The Gemini API key is stored in Odoo ir.config_parameter.

Recommendations:

• Limit Settings access to trusted administrators.
• Do not paste API keys into logs, chat messages, tickets, or screenshots.
• Rotate keys if they are exposed.
• Use Google Cloud restrictions where possible.

Error Sanitization

The connector sanitizes common API key patterns in error messages before showing them to users.

Still, administrators should avoid sharing full tracebacks publicly.

Test Connection Fails

Check:

• API key is saved.
• Model is available for the current project.
• API base URL is correct.
• Google AI Studio quota is available.
• Odoo server can reach the internet.

Quota Error

If Google says quota is exceeded:

• Check Google AI Studio Rate Limit page.
• Try another Gemini model with available quota.
• Wait for daily reset.
• Enable billing if production usage is needed.

Module Does Not Appear in Settings

Check:

• Module is installed.
• Browser page is refreshed.
• User has access to General Settings.
• Odoo assets are refreshed if needed.

Command:

python3 ~/odoo/odoo-18/odoo-server \
   -c ~/odoo/conf/odoo.conf \
   -d your_database \
   -u sm_gemini_connector \
   --stop-after-init \
   --no-http

Python Model

sm.gemini.connector.service

Important Methods

call_ai_chat(messages, temperature=0.7, max_tokens=1000, model=None)
call_gemini_chat(messages, temperature=0.7, max_tokens=1000, model=None)
test_gemini_connection()
parse_json_response(response)

Settings Fields

sm_gemini_connector_gemini_api_key
sm_gemini_connector_gemini_model
sm_gemini_connector_gemini_custom_model
sm_gemini_connector_gemini_api_base_url
sm_gemini_connector_request_timeout
sm_gemini_connector_tokens_used
sm_gemini_connector_monthly_token_limit

Compatibility

• Tested on Odoo 18.0 Community & Enterprise.
• Requires Odoo base and web.
• Uses standard Odoo ORM, res.config.settings, and ir.config_parameter.
• Uses Python requests library.