Google Gemini

🔑 BYO Key — AI

Generate AI responses and images with Google Gemini. Chat, content generation, analysis, reasoning, and AI image generation. BYOK — bring your own Google AI API key. Call list-models for the models your key can actually reach; Google retires model names regularly.

🔑 Setup Required — Bring Your Own Key This integration requires API credentials. Store them once via setup_integration or the Dashboard Vault — they're encrypted with AES-256-GCM and never exposed.

Required Credentials

gemini_api_key

API Key

Google AI API key from aistudio.google.com/apikey. New keys are auth keys; older AIza keys are being phased out by Google (rejected from September 2026).

Get your API key at ai.google.dev →

Endpoints (3)

POST generate-content

Generate content with Gemini. Send a text prompt and receive an AI response. There is no default model: pass one, and use list-models to see which ones your key can reach.

Field	Type	Required	Description
`model`	string	✓ Yes	Gemini model id. Call list-models first — Google retires names and limits older ones to existing users. Current general-purpose choice at time of writing: "gemini-3.5-flash-lite" (fast, cheap) or "gemini-3.8-flash" (most capable). Models from the 2.x range are no longer available to new users.
`contents`	array	✓ Yes	Conversation contents. Example: [{"parts": [{"text": "Explain quantum computing"}]}].

POST generate-image

Generate images from text prompts using Gemini image models. Returns inline image data. Replaces the deprecated Imagen 3 integration. Send your prompt in contents and set generationConfig.responseModalities to ["IMAGE"]. Pass an image-capable model id — call list-models to see which ones your key can reach.

Field	Type	Required	Description
`model`	string	✓ Yes	Image-capable Gemini model id, required. At time of writing: "gemini-3.1-flash-image" (fast) or "gemini-3-pro-image" (higher quality). Call list-models for what your key can reach. Do NOT use imagen-3.0 or 2.x models — retired.
`contents`	array	✓ Yes	Prompt as contents array. Example: [{"parts": [{"text": "A professional blog header showing a modern Dutch office with warm lighting"}]}].
`generationConfig`	object	No	Generation settings. MUST include responseModalities: ["IMAGE"] for image output. Optional: aspectRatio ("1:1","3:4","4:3","9:16","16:9"), numberOfImages (1-4).

GET list-models

List available Gemini models with their capabilities, input/output token limits, and supported methods.

No input parameters required.

MCP Tool Names

When using this integration through an AI assistant (Claude, ChatGPT, Cursor, etc.), the endpoints are available as MCP tools:

Endpoint	MCP Tool Name
generate-content	`gemini_generate_content`
generate-image	`gemini_generate_image`
list-models	`gemini_list_models`

← Back to all integrations