openclaw/docs/providers/pollinations.md

summary	read_when
Use Pollinations.ai unified Gen-AI platform with Clawdbot	You want to use Pollinations.ai models in Clawdbot You need access to multiple AI providers through a single API

Pollinations

Pollinations.ai is an open-source Gen-AI platform that provides a unified API gateway for accessing multiple AI models including GPT-5, Claude, Gemini, Flux (image generation), and more. It uses an OpenAI-compatible API format, making it easy to integrate with Clawdbot.

Features

• Unified API — Access GPT-5, Claude, Gemini, Grok, and more through a single endpoint
• Image Generation — Flux, GPT Image, Seedream, and other image models
• Video Generation — Text-to-video with Seedance and Veo (alpha)
• Audio — Text-to-speech, speech-to-text, and voice options
• Pay-as-you-go — Simple Pollen credits system ($1 ≈ 1 Pollen)
• OpenAI-compatible — Works with existing OpenAI API clients

Getting your API key

1. Visit enter.pollinations.ai
2. Log in with your GitHub account
3. Create an API key from the dashboard
4. Choose between:
   • Publishable keys (pk_) — Frontend use, with rate limits
   • Secret keys (sk_) — Server-side only, no rate limits (keep secret!)

You can scope each API key to specific models or allow access to all models.

Configuration

CLI setup

clawdbot onboard --auth-choice pollinations
# or non-interactive
clawdbot config set env.POLLINATIONS_API_KEY "sk-..."

Config snippet

{
  env: { POLLINATIONS_API_KEY: "sk-..." },
  models: {
    providers: {
      pollinations: {
        baseUrl: "https://gen.pollinations.ai/v1",
        apiKey: "${POLLINATIONS_API_KEY}",
        api: "openai-completions",
        models: [
          {
            id: "openai",
            name: "GPT-5 via Pollinations",
            reasoning: false,
            input: ["text", "image"],
            cost: { input: 0.002, output: 0.006, cacheRead: 0.001, cacheWrite: 0.0025 },
            contextWindow: 128000,
            maxTokens: 16384
          },
          {
            id: "claude",
            name: "Claude via Pollinations",
            reasoning: false,
            input: ["text", "image"],
            cost: { input: 0.003, output: 0.015, cacheRead: 0.0003, cacheWrite: 0.00375 },
            contextWindow: 200000,
            maxTokens: 8192
          },
          {
            id: "gemini",
            name: "Gemini via Pollinations",
            reasoning: false,
            input: ["text", "image"],
            cost: { input: 0.001, output: 0.002, cacheRead: 0.0001, cacheWrite: 0.0005 },
            contextWindow: 128000,
            maxTokens: 8192
          }
        ]
      }
    }
  },
  agents: {
    defaults: {
      model: { primary: "pollinations/openai" }
    }
  }
}

Available models

Pollinations provides access to many models through its unified API:

Text Generation

• openai — GPT-5
• claude — Claude (latest)
• gemini — Gemini Pro
• gemini-large — Gemini with code execution
• gemini-search — Gemini with Google Search
• grok — Grok
• nomnom (alias gemini-scrape) — Web research and scraping

Image Generation

• flux — FLUX models for text-to-image
• gptimage-large — GPT Image
• seedream — Seedream image generation
• klein — FLUX.2 klein for fast text-to-image

Video Generation (alpha)

• seedance — Text-to-video
• veo — Text-to-video

Authentication

Pollinations supports two authentication methods:

1. API Key in URL (for simple requests):

   curl 'https://gen.pollinations.ai/text/hello?key=YOUR_API_KEY'
   

2. Bearer Token in Header (recommended for Clawdbot):

   curl 'https://gen.pollinations.ai/v1/chat/completions' \
     -H 'Authorization: Bearer YOUR_API_KEY' \
     -H 'Content-Type: application/json' \
     -d '{"model": "openai", "messages": [{"role": "user", "content": "Hello"}]}'
   

Clawdbot uses the Bearer Token method automatically when you set POLLINATIONS_API_KEY.

Security

• Never share your API key publicly — Don't commit keys to Git repositories
• Use environment variables — Store your key in .env files
• Use secret keys for production — Secret keys (sk_) have no rate limits and should only be used server-side
• Scope your keys — When creating keys, restrict them to only the models you need

Pricing

Pollinations uses a Pollen credits system:

• $1 ≈ 1 Pollen
• Pay-as-you-go billing
• Different tier levels (Seed, Flower, Nectar) provide different usage quotas
• Seed tier is automatic on first login

Check your balance and usage:

# Balance
curl 'https://gen.pollinations.ai/account/balance' \
  -H 'Authorization: Bearer YOUR_API_KEY'

# Usage history
curl 'https://gen.pollinations.ai/account/usage' \
  -H 'Authorization: Bearer YOUR_API_KEY'

Notes

• OpenAI-compatible — Pollinations uses the OpenAI Completions API format, so it works seamlessly with Clawdbot's OpenAI provider adapter
• Multi-modal support — Many models support both text and image inputs
• Tool support — Gemini models include code_execution and google_search tools
• Response formats — Some responses may include special content blocks (image_url, thinking) depending on the model
• Beta status — Pollinations is in active development; features and pricing may change

Troubleshooting

Authentication errors

If you see 401 Unauthorized:

• Verify your API key is correct
• Make sure you're using the right key type (secret keys for server-side)
• Check that your key hasn't expired

Model not found

If a model isn't available:

• Check the API documentation for current model names
• Some models may be in limited access or beta
• Try using a different model from the list above

Rate limits (publishable keys only)

Publishable keys (pk_) have rate limits. If you hit limits:

• Use a secret key (sk_) for server-side usage
• Upgrade your tier for higher quotas
• Check your usage at enter.pollinations.ai

See Also

• Pollinations.ai Website
• API Documentation
• Model Providers - Overview of all providers
• Model Selection - How to choose models
• Configuration - Full config reference