Getting Started with Pollinations AI
If you have fifteen minutes, you can leave with a generated image and a text task completed — then come back when you care about models, keys, and API details.
Quick Start - No Signup Required
Pick a tool, write a plain prompt, hit generate — everything else on this page is optional depth
Choose Your Tool
Visit our web interface and select what you want to create:
- Image Generator - Create images with Flux, GPT Image, Seedream
- Text Generator - Generate text with GPT-5, Claude, Gemini
Enter Your Prompt
Describe what you want to create. Be specific for best results:
Image Example: "A beautiful sunset over mountains, digital art style, vibrant colors"
Text Example: "Write a 500-word blog post about AI image generation"
Generate & download
Hit generate, wait out the spinner, then save what worked — if it is close but wrong, change one line of the prompt and try again before you switch models.
API integration
From browser trial to curl — the same endpoint you clicked around in the UI
Step 1: Get your API key (optional)
Grab a key from enter.pollinations.ai when you leave prototype territory — casual tries in the browser often work without one.
Step 2: Use the Unified API
All Pollinations AI services are accessible through the unified endpoint:
https://gen.pollinations.ai
Step 3: Make Your First API Call
# Image Generation
curl 'https://gen.pollinations.ai/image/a%20beautiful%20sunset' -o image.jpg
# Text Generation
curl 'https://gen.pollinations.ai/text/Hello%20world'
# With API Key
curl 'https://gen.pollinations.ai/image/a%20cat?key=YOUR_KEY'
For detailed API documentation, visit our API documentation page.
Choosing a model
Rules of thumb — swap if quality plateaus after a few prompt tweaks
For Images
- Flux: Default for “make me a picture” — good first try
- GPT Image Large: When you need crisp detail and fewer mushy edges
- Seedream: When you want a more stylized or painterly read
For Text
- GPT-5: General writing and reasoning
- Claude: Long explanations and careful edits
- Gemini: When tooling/search routes are enabled
- Qwen3-Coder: Scripts and refactors
Tips
- Start plain, then add one constraint per iteration
- Name the medium: photo, illustration, 3D render…
- If two models disagree, your prompt might be underspecified
- See trade-offs on the models page
Additional resources
Deeper docs when the quick path is not enough
Enough reading — try a tool
Image and text workspaces are one click away; come back to this page when you wire up the API