Version 1.30.0
9 days ago by Tech admin
The release includes:
- New Archive Mode: You can now archive core resources like datasets, policies, guardrails, agents, and dashboards. Archiving acts as a soft delete, preserving the resource and its configuration, but preventing further use or modifications. Unlike deletion, archiving is reversible. Users can view active resources by default or switch to view archived ones. Archived resources aren’t counted towards your tier limit.
- Bulk Operation: Need a major cleanup? Tired of deleting or archiving resources one by one? You can now select multiple resources and perform bulk deletes or archives. No more excuses for a messy account!
- Streaming: For a more engaging experience, we’ve added a streaming option to our deployed agents, complementing the existing batch mode. Users can see each action the agent takes live, like checking input validity with guardrails or using tools for answers. Output is streamed in chunks, and our iframe uses this streaming endpoint by default.
- Guardrail Violation: Managing multiple guardrails in your agent? Some might be critical enough to block agent responses, while others are for auditing. You can now control these settings at the individual guardrail level.
- Commercial Tiers: We’ve aligned our platform with a new 5-tier model to better serve our business needs. We’ve also improved the upgrade and downgrade experience. Users now see only relevant upgrade options, and if downgrading (explicitly or due to payment issues), the platform will automatically archive excess resources to keep users within their new tier limits.
- Other Updates:
- Our SDK now exposes guardrail check results when using the agent ask API.
- Increased the size of supported knowledge files to 512MB from 10MB.
- Several production bug fixes.
- Embedded Google tags for better campaign tracking.