Navigate to MCP Gateway → Dashboard to view the MCP server status, active providers, tool statistics, and quick-access buttons for generating API keys and testing tools — all from a single screen.

Manage all AI-callable actions from MCP Gateway → Configuration → AI Tools. Each tool supports built-in ORM operations, decorated Odoo methods, or custom Python snippets, with configurable provider, model, and consent settings per tool.

The Odoo MCP Gateway provides pre-configured access to core Odoo models for AI assistants, with the flexibility to enable, disable, or add additional models based on project requirements.

Go to MCP Gateway → Configuration → Providers → New. Enter your provider name (e.g. OpenAI, Anthropic, Google Gemini, or Ollama), paste your API key, and set a priority. Multiple providers can be active simultaneously — the gateway picks the highest-priority one automatically with fallback support.

After saving a provider, click Fetch Models to auto-import all available models from the provider API. Then mark your preferred model as Default for Use. For Anthropic, models are added manually — use the exact model IDs such as claude-sonnet-4-6 or claude-opus-4-7.

In the AI Hub Dashboard, click New MCP Key. Select the target platform (Claude Desktop, Claude Code CLI, claude.ai, or OpenAI Codex CLI). The wizard generates a secure API key and renders the exact configuration snippet to paste into your AI client. The key is shown only once — copy it immediately.

Generate an MCP key and select Claude Code CLI as the platform. The wizard renders a ready-to-run shell command. Open a terminal, paste the command, and the MCP server is registered automatically: claude mcp add-json odoo-ai-hub '{...}' Run claude mcp list to verify registration, then restart Claude Code and type /mcp in a session to confirm the odoo-ai-hub server is connected with all Odoo tools available — no file editing required.

Generate an MCP key and select Claude Cloud (claude.ai) as the platform. The wizard produces a single URL with the API key embedded: https://your-odoo-domain.com/mcp_gateway?api_key=YOUR_KEY In claude.ai go to Settings → Integrations → Add custom integration, paste the URL, and click Save. Start a new conversation — Odoo tools such as search_records, create_record, analyze_records and ask_ai will be available immediately. Your Odoo server must be reachable over public HTTPS.

Go to MCP Gateway → Tools to see all registered AI tools. Each tool shows its name, description, and auto-generated JSON schema. Click Test Tool on any entry to send live parameters and inspect the real output — useful for verifying your provider and model are correctly configured before connecting an external AI client.

Navigate to MCP Gateway → Bot Channels → New. Select a platform, paste your bot token (from @BotFather for Telegram, Meta Developer Console for WhatsApp, or Discord Developer Portal for Discord), then click Connect. Odoo registers the webhook automatically and shows a direct link to your live bot. The Web Widget requires no external account — just create the channel and send POST requests to the generated endpoint.

Go to MCP Gateway → Tool Logs to audit every AI tool call made through the gateway. Each log entry records the tool name, input parameters, provider used, model used, execution duration, and the result. Use this view to debug unexpected outputs, track usage costs, and verify that tools are calling the correct AI model.

Go to Settings → MCP Gateway to generate the Webhook Secret (appended to every bot webhook URL to reject unauthorized requests) and the Bot MCP API Key (used internally by the bot gateway to authenticate tool calls). Both values are auto-generated with a single click and stored securely in Odoo's system parameters.