| name | pollinations |
| description | Use when adding Pollinations image, text, audio, video, or embedding features to this app, or when updating its AI-generation flow. |
Pollinations integration
Pollinations provides production image, text, audio, video, and embedding inference. Use the live API documentation at https://gen.pollinations.ai/docs and model discovery endpoints instead of hardcoding a model catalog.
Security
- Keep
sk_ keys server-side in Replit Secrets. Never put them in browser code, client bundles, commits, or logs.
- Use a small server route or approved serverless function for requests needing a secret key.
- Validate user input, return generic upstream errors, and avoid logging prompts or credentials by default.
This starter
POST /api/images accepts { "prompt": "..." } and proxies the image request.POLLINATIONS_API_KEY must be configured in Replit Secrets before image generation works.- For new modalities, check the current docs first, then add a narrowly scoped server route and a visible user-facing error state.
Useful endpoints
- Image:
GET https://gen.pollinations.ai/image/{prompt}
- OpenAI-compatible text:
POST https://gen.pollinations.ai/v1/chat/completions
- Current text models:
GET https://gen.pollinations.ai/v1/models
