Replicate
inference_host
65 signals tracked
Cog Playground: New Video and Image Rendering Features
Added video support for before/after slider and pixelated image rendering with size controls
7 Nov 2025
HighCapabilityReplicate: Deployment failure notifications and customizable timeouts for Enterprise
Enterprise plan customers can now receive automatic email notifications when deployments fail during setup and configure custom timeout periods for model initialization. These features help catch deployment issues earlier and accommodate models requiring longer setup times for downloading large files or loading weights. Both features can be configured in deployment settings.
3 Nov 2025
MediumCapabilityReplicate launches Python SDK public beta with HTTP API support
Replicate has released a public beta of their Python SDK, providing full support for all HTTP API operations to simplify AI model integration. The update also includes improved playground handling for different output types and streaming, plus documentation search shortcut fixes. This is a new capability enhancement that requires no immediate action from existing users.
24 Oct 2025
InfoCapabilityReplicate introduces Cancel-After header for automatic prediction cancellation
Replicate introduced a Cancel-After header that allows users to set deadlines for predictions to automatically cancel if they don't complete within a specified duration. This feature is designed for real-time applications and users are only charged for canceled predictions, not aborted ones. No action required as this is an optional new feature.
16 Oct 2025
MediumCapabilityReplicate platform performance and UI updates released
Replicate has implemented several performance enhancements including faster overall platform speed, improved API filtering, and updated dashboard navigation components. These changes also include minor UI improvements like making the Official label clickable on model pages. No action required from users as these are backend improvements and interface enhancements.
10 Oct 2025
MediumCapabilityReplicate API adds sorting to /v1/models endpoint
The models API endpoint now supports sort_by and sort_direction query parameters for easier retrieval of newest models. Users can sort by model creation date or latest version date in ascending or descending order. This is an additive enhancement that doesn't require any changes to existing integrations.
8 Oct 2025
LowCapabilityBilling: PDF Invoice Download Added
Users can now download PDF invoices directly from billing settings for both prepaid credit and monthly billing options. This new feature provides easier access to billing documentation without requiring external navigation to Stripe for monthly users. No action required - this is an enhancement to existing billing functionality.
8 Oct 2025
LowCapabilityReplicate: New API endpoint for updating model properties via PATCH
Replicate has added a new API endpoint that allows users to update model properties using PATCH requests to /v1/models/{owner}/{name}. Users can now programmatically update model descriptions, README content, GitHub URLs, paper URLs, weights URLs, and license URLs. This is a new feature that enhances API functionality without requiring any changes to existing implementations.
6 Oct 2025
MediumCapabilityReplicate launches new search API and improves playground UI for image handling
Replicate has introduced a beta search API for finding models, collections, and documentation in a single call. The platform also improved the playground interface to better handle output images as inputs and enhanced homepage performance. These are feature additions that don't require any action from existing users.
26 Sept 2025
InfoCapabilityReplicate adds new search API for models, collections, and documentation
Replicate has introduced a new search API that enables easier discovery of models, collections, and docs. The API is available in TypeScript and Python SDKs as well as MCP servers. This is an additive feature that enhances the platform's usability without requiring any immediate action from existing users.
16 Sept 2025
LowCapabilityReplicate adds invoices, torch compile caching, and web URLs for predictions
Replicate introduced three new platform features: invoicing for prepaid credit purchases, torch compile caching that speeds up model startup by 2-3x, and web URLs in prediction API responses for browser viewing. These are additive improvements that enhance platform functionality without requiring any changes to existing integrations.
12 Sept 2025
LowCapabilityReplicate: torch.compile caching improves Flux model startup times
Replicate has implemented caching of torch.compile artifacts to reduce model startup times by 2-3x for Flux models while maintaining 30%+ inference speed improvements. This is an automatic optimization that benefits existing torch.compile users without requiring any changes to their implementation.
8 Sept 2025
InfoCapabilityReplicate enhances platform UX with Artificial Analysis and email verification
Replicate enhanced their platform with new Artificial Analysis arena rankings in search results and added email verification for organization signups. Additional improvements include better billing dashboard rendering and general site navigation enhancements. These are informational updates that improve user experience without requiring any action from existing users.
29 Aug 2025
InfoCapabilityReplicate UI improvements: model pages, homepage, predictions, and search
Replicate has made several user interface enhancements including overhauled model page headers, updated homepage with fresh model data, clearer predictions filtering, and improved search results. These changes also include bug fixes for 404 errors on collections and video model display issues. No action required from users as these are interface improvements.
14 Aug 2025
LowCapabilityReplicate Unified POST /v1/predictions Endpoint for All Model Types
Replicate has enhanced the POST /v1/predictions endpoint to work with both official and community models, removing confusion about endpoint selection. The change is fully backward compatible and existing code will continue working without modifications. Users can now use a single endpoint for all models while retaining access to the legacy model-specific endpoints.
5 Aug 2025
MediumCapabilityReplicate: New Search, Enterprise Page, and Playground Filtering
Replicate has introduced a new search experience platform-wide, added an enterprise page, and enabled model filtering in the playground. Several bug fixes were also implemented including fixes for dark mode persistence and model creation failures. These are enhancement updates that improve user experience without requiring any action from users.
1 Aug 2025
LowCapabilityReplicate introduces new Cog production runtime component with Go
Replicate has introduced a new production runtime component for Cog that handles predictor schema validation, prediction execution and HTTP serving. Model authors can test the new runtime by upgrading to Cog >= 0.16.0 and adding build.cog_runtime: true to cog.yaml. This runtime will become the default in a future release, after which the existing runtime will be deprecated.
21 Jul 2025
MediumCapabilityReplicate platform improvements: model creation, video covers, UI updates
Replicate has streamlined the model creation process and added support for video cover images on models. Additional UI improvements include better organization of API tokens and clearer invoice formatting for small charges. These are enhancement updates that don't require user action.
18 Jul 2025
LowCapabilityReplicate introduces prepaid credit billing for new accounts
New accounts are now billed through prepaid credit instead of monthly billing as of July 16, 2025. Existing monthly billing users can continue unchanged for now, but Replicate plans to migrate most accounts to prepaid credit in the future. Users who want to switch early can contact support or manage billing through their account settings.
16 Jul 2025
HighPricingReplicate: Documentation, API, and Playground Updates
Replicate updated model documentation to clarify required vs optional inputs, launched environment variables in running containers, and improved API schema completeness. Additionally fixed a UI issue in the API playground related to image uploads. These are informational improvements that enhance usability without requiring user action.
4 Jul 2025
LowCapability
Get alerts for Replicate
Never miss a breaking change. SignalBreak monitors Replicate and dozens of other AI providers in real time.
Sign up free — no credit card required