Replicate
ModelCloud API for running and fine-tuning AI models
- Price
- Pay per use
- Access
- API token
About
Cloud platform for running public AI models and deploying your own models packaged with Cog. Public models are billed per output, per token or per second of GPU time. Select models can be tried for free before you set up billing.
What you can do with it
- Run public models for images, video, text, virtual try-on, lip sync and photo restoration
- Fine-tune an image model like FLUX on your own photos
- Deploy a custom model packaged with Cog and scale it on GPUs
Get started
- Create an API token
- Export it as REPLICATE_API_TOKEN
Example
curl -s -X POST -H 'Prefer: wait' \
-H "Authorization: Bearer $REPLICATE_API_TOKEN" \
-H 'Content-Type: application/json' \
-d '{"input": {"prompt": "A photo of a bear riding a bicycle over the moon"}}' \
https://api.replicate.com/v1/models/black-forest-labs/flux-schnell/predictionsDetails
- Hosting
- Hosted service
- Available in
- Worldwide
- Official SDKs
- Python, JavaScript/TypeScript, Go, Swift
- MCP server
- Remote