Blog
Releases and updates.
What shipped and why. The full history lives in the changelog.
AgentUse v0.21: providers as plugins and a judge that remembers
Model providers install as plugins from a reviewed shortlist, agent creation moves into a draft-and-refine workspace, and pre-review judges keep one session across attempts.
AgentUse v0.20.0: revise an agent from the run that went wrong
Completed and paused runs can propose a reviewed change to their own agent, and the dashboard gains a command palette and live-work navigation.
AgentUse v0.19: a native Mac app and learnings that earn their place
AgentUse for Mac bundles the runtime and dashboard in a menu-bar app, learnings are vetted before they take effect, and providers and schedules are managed from the dashboard.
AgentUse v0.18.0: smaller agents and calmer operations
Write focused agent instructions and review live work in a redesigned serve UI built for calmer daily operations.
A field guide to agentuse test
Tuning an agent used to mean real runs with real consequences. How agentuse test turns that into a safe loop: every approval branch, a rogue probe, and a CI gate.
AgentUse v0.17.0: model aliases, agentuse test, and learnings that graduate
Stop pinning model versions, validate agents end-to-end without side effects, and let proven corrections become permanent instructions. Plus: restarting serve no longer kills running agents.