Changelog

Web

• Launched official models
• Added latest predictions to the deployment detail page
• Fixed an issue with displaying certain pages on Replicate when logged out
• Updated the model detail API playground to make it possible to train with inputs that aren’t .zip
• Made it clearer when creating a deployment that flux fine-tunes shouldn’t be deployed
• Fixed a bug where numeric inputs could increment on scroll on the model detail API playground
• Fixed a bug that could cause inaccessible models to show up in collections
• Improved the experience of sharing predictions via web
• Improved clarity of autoscaling charts for deployments
• Made it easier to copy the model identifier (for official models) or the model identifier and sha (for community models)
• Updated chart for GPU metrics to make values easier to differentiate

Playground

• Added an icon to thumbnails to differentiate video content from other types
• Made it easier to see when errors occur and why
• Added the ability to remove multiple items from the grid on the playground
• Fixed a bug when selecting individual items that could cause all items to be deselected
• Added Vercel AI SDK snippets to playground
• Improved the loading and handling of large number of videos

Docs

• Updated deployment documentation
• Improved syntax highlighting across docs

Official models are always on and have predictable pricing.

A collection of updates to Replicate for the week ending January 17, 2025.

A collection of updates to Replicate for the week ending December 20, 2024.

We've stopped our experimental support for language model training

You can now monitor how much GPU memory your deployments are using so you can make sure you’re using your GPUs efficiently.

A collection of updates to Replicate for the week ending December 06, 2024.

A collection of updates to Replicate for the week ending November 22, 2024.

A collection of updates to Replicate for the week ending November 8, 2024.

Python 3.7 model support ends on November 15th.

A collection of updates to Replicate for the week ending October 25, 2024.

Finally. You can switch it in the footer.

Playground is a way to quickly try out and compare the output of models on Replicate.

Documentation now has a cleaner design, better navigation, and lots of new content to help you build with Replicate.

Our client libraries and API are now much faster at running models, particularly if a file is being returned.

Large logs may be truncated to manage platform load.

Cleaning old predictions to improve API performance and speed.

The training detail page now includes a JSON metadata tab.

Programmatically search for public models using the API

You no longer need to pass `"stream": true` with your prediction request to get back a stream URL - if the model supports streaming, you'll always get one back.

Securely pass sensitive values to models.

Disable API tokens from the web to prevent unauthorized use.

We publish automated feeds about product updates and platform incidents.

Delete models, versions, and deployments via web or HTTP API.

Validate incoming webhooks so you know they're really coming from Replicate.

You can now search for your existing deployments on the website.

We've added new UI features for viewing and tracking webhooks.

The API now validates JSON request body fields.

T4 models now get up to 16 GB of RAM, with no changes to existing prices.

Store API tokens in config, not code, for security.

Bearer token used in HTTP Authorization header for access.

Manage models with more control using our new deployments feature.

We've added a web-based viewer for 3D model outputs.

We've updated our model collections to be more task-oriented.

Use webhooks to receive real-time updates about your predictions and trainings.

Create custom SDXL image generation models without coding.

New slider for comparing image inputs to outputs.

Code snippets for various languages now available on models.

You can now create models programmatically using the API.

Improved training pages for fine-tuning models.

View prediction parameters as JSON on prediction detail page

New API endpoint for listing public models.

Create deployments for controlled model running and customization.

Prediction UUID added as query parameter on web refresh.

View training logs in full-screen mode.

Browser tab favicon shows prediction status.

Our API now supports live language model output with SSE streams.

Create multiple personal API tokens for your user account.

NVIDIA A40 GPUs now supported.

Model pages now show training hardware type and cost.

We've launched a fine-tuning API for training custom language models.

See Git commits and tags in model versions on Replicate.

Website download mechanism improved for multiple and single outputs.

View detailed invoice summaries with cost and model breakdowns.

We've published a Swift client library for AI-powered app development.

Collaborate with your team on Replicate using organizations.

We're released a Node.js client library for model predictions.

The "get a model" API now returns more metadata.

All models have OpenAPI schema objects for their inputs and outputs.

HTTP API reference is now also available as a structured OpenAPI JSON schema.

Browse models on our new Explore page.

API sends webhook events at different prediction lifecycle stages.

API now provides complete Python example code snippets.

Cancel predictions even after navigating away from the page.

Install Cog on macOS with Homebrew using brew command.

DreamBooth API now supports img2img with prompt and image input.

Delete prediction button available on prediction detail page

API prediction data is automatically removed after one hour.

We are now publishing a dedicated changelog for product updates.

Stable Diffusion now has release notes for version changes.

Increased default rate limits for predictions up to 600 per second.

Infrastructure improvements increase reliability and speed.

Run models on Nvidia A100s via hardware upgrade option

Set monthly spend limit to avoid surprise bills easily

API now supports webhooks for prediction completion notifications

Curated collections of models for similar tasks, starting with style transfer.

Scrub through model predictions to see how they evolved.