Model Context Protocol

- 🔓 100% Open Source — code, decisions, roadmap all public
- 🤝 Community-Built — 500+ projects already using our APIs
- 🌱 Pollen Quests — earn Pollen by completing Quests (in alpha)
- 🖼️ Image Generation — Flux, GPT Image, Seedream, Kontext
- 🎬 Video Generation — Seedance, Veo (alpha)
- 🎵 Audio — Text-to-speech and speech-to-text
- 🎣 _Easy-to-use Packages_ (Packages)

<a href="https://star-history.com/#pollinations/pollinations&Date">
<picture>
<source media="(prefers-color-scheme: dark)" srcset="https://api.star-history.com/svg?repos=pollinations/pollinations&type=Date&theme=dark" width="600" />
<source media="(prefers-color-scheme: light)" srcset="https://api.star-history.com/svg?repos=pollinations/pollinations&type=Date" width="600" />

</picture>
</a>

Add this to your MCP client configuration:

{
  "mcpServers": {
    "pollinations": {
      "command": "npx",
      "args": ["@pollinations/mcp"]
    }
  }
}

1️⃣ Get your API key
Sign up at enter.pollinations.ai to generate your key.

2️⃣ Choose what you want to generate
Pollinations supports:
- 🖼 Images
- 📝 Text
- 🔊 Audio
- 🎬 Video

3️⃣ Make your first request
Use one of the examples below to generate your first AI output in seconds.

npx @pollinations/mcp

Community alternatives like MCPollinations and Sequa MCP Server are also available.

AI assistants can:

- Generate images from text descriptions
- Create text-to-speech audio with various voice options
- Play audio responses through the system speakers
- Access all pollinations.ai models and services
- List available models, voices, and capabilities

For more advanced usage, check out our API documentation.

Get your API key at enter.pollinations.ai

- generateImageUrl: Generate an image URL from a text prompt. Returns a shareable/embeddable URL without downloading the image. Supports all image models and parameters including image-to-image with the 'image' parameter.
- generateImage: Generate an image from a text prompt and return the base64-encoded image data. Full control over all generation parameters. Supports image-to-image with kontext, seedream, nanobanana models.
- generateImageBatch: Generate multiple images in parallel (up to 10). Best with sk_ keys (no rate limits). pk_ keys will be rate-limited. Each prompt generates one image with shared parameters.
- generateVideo: Generate a video from a text prompt or image. Models: veo (text-to-video, 4-8s, audio), seedance (text/image-to-video, 2-10s), seedance-pro (best quality). Use 'image' parameter with seedance models for image-to-video generation.
- generateVideoUrl: Generate a video URL from a text prompt. Returns a shareable/embeddable URL without downloading the video. Models: veo (4-8s, audio), seedance (2-10s), seedance-pro (best quality).
- describeImage: Analyze and describe an image using vision-capable AI models. Pass an image URL and optionally a custom prompt. Great for image captioning, content analysis, OCR, and visual Q&A.
- analyzeVideo: Analyze a video using AI. Supports YouTube URLs and direct video links. Uses gemini-large for native video understanding (frames + audio). Great for video summarization, content analysis, and Q&A about videos.
- listImageModels: List all available image and video generation models with their capabilities. Shows which models support image-to-image, video generation, and other features. Models are fetched dynamically from the API.
- generateText: Generate text from a simple prompt. Easy-to-use text generation with essential parameters. For advanced features (tool calling, multi-turn, audio), use chatCompletion instead.
- chatCompletion: OpenAI-compatible chat completions with ALL parameters. Supports:
- Multi-turn conversations with message history
- Function/tool calling for AI agents
- Audio input/output (openai-audio model)
- Reasoning mode (kimi, perplexity-reasoning, openai-large, gemini-large)
- JSON/structured output
- Built-in Gemini tools (google_search, code_execution, etc.)
- Perplexity web search with citations
- listTextModels: List all available text generation models with their capabilities. Shows which models support reasoning, tools, audio, vision, etc. Models are fetched dynamically from the API.
- webSearch: Search the web for real-time information using search-enabled models. Returns answers with source citations. Great for current events, facts, research.
- getPricing: Get pricing information for all models. Shows cost per token/image in pollen.
- respondAudio: Generate an audio response to a text prompt. The AI will respond to your prompt with speech.
- sayText: Generate speech that says the provided text verbatim. Direct text-to-speech.
- listAudioVoices: List all available audio voices and supported formats. Voices are fetched dynamically from the API.
- transcribeAudio: Transcribe audio from a URL. Uses gemini-large for accurate speech-to-text transcription.
- setApiKey: Set your pollinations.ai API key for authenticated requests. Get your key at https://enter.pollinations.ai/keys
- getKeyInfo: Get information about the currently set API key
- clearApiKey: Clear the stored API key
- getBalance: Get the current Pollen balance for the authenticated API key. Returns key-scoped balance if the key has its own budget, otherwise account-wide. Requires an API key with 'account:usage' permission.
- getUsage: Get usage history for the authenticated key. By default returns the most recent per-request records (model, cost, tokens, latency). Set 'daily: true' for a daily aggregated summary grouped by date and model. Requires 'account:usage' permission.
- listQuests: List quests and earned rewards for the authenticated account. Read-only; claiming rewards requires a dashboard session. Requires 'account:usage' permission.