Changelog
See our latest feature releases, product improvements and bug fixes
Enhanced API key management
We've shipped several small quality-of-life improvements for managing API keys when creating workspace keys for specific models. You can now filter the list, making it easier to find the right key,...
Aug 16, 2024Improved filtering by Chainlet
We've heard your feedback! You can now easily see which chainlet you're filtering by when viewing logs and metrics. We've also improved the filter popover, making it simpler to navigate and filter in...
Aug 15, 2024Async API endpoint code example
Interested in running async inference ? We've made it easier to get started by adding a code example in the new API endpoint modal. Just click API in the top right of the page and toggle to Async API...
Aug 14, 2024Improved replica selector
To handle massive bursts of traffic, we've helped user scale to hundreds of replicas on model deployments. To make it easier to set minimums and maximums for large numbers of replicas, we've added...
Jul 9, 2024Pin frequently-used models for easy access
By popular demand: you can now pin models and chains to the top of your workspace! To pin an item, click into the ... menu on any model or chain and select Pin . You can pin up to 6 items for quick...
Jul 2, 2024Export billing usage data
You can now export your model and billing usage data for in-depth analysis in your preferred tool. The exported CSV includes all the model usage data from the selected billing period. This includes:...
Jun 11, 2024Run model inference asynchronously on Baseten
We’re thrilled to announce that you can now run async inference on Baseten models! This unlocks some powerful inference use cases: Scalable processing: Schedule tens of thousands of inference...
Apr 22, 2024Refreshed model overview page with deployment statuses
We’ve revised the model overview page to give you more information about model deployments at a glance. Each model card now contains: The total number of deployments. Counts for deployment statuses:...
Apr 3, 2024Improved log filtering
You can now filter logs through the main text input. Just start typing the filter you’re looking for, like level , and autocomplete options will appear. Currently, logs filter by: Log level: zoom in...
Mar 26, 2024Permit inference on unhealthy models
A model enters an “unhealthy” state when the deployment is active but there are runtime errors such as downtime on an external dependency. We now permit inference requests to proceed even when a...