Deployment reviews let LatchLoop inspect a successful preview deployment in a real browser and report whether the deployed app appears to match the task.
LatchLoop can open the preview, interact with the UI, capture evidence, and summarize whether the task-specific goal looks complete.
Deployment review is triggered after LatchLoop receives a successful deployment event for a task.
If a newer deployment review supersedes an older one, the older run is marked as superseded so the task focuses on the latest evidence.
For each project, you can configure:
Use these settings for project-wide expectations, such as “verify the dashboard loads without console errors” or “start all reviews at /app.”
Sometimes a browser review would be misleading. For example, the task may require a database migration, edge function deploy, seed data, or a backend rollout before the UI can be tested correctly.
In those cases, LatchLoop can mark the review as blocked before browser launch and show the reason in the task UI.
You can manually rerun the deployment review later after the prerequisite is complete.
Deployment review results can include:
If a review fails, LatchLoop can use the review findings to focus a follow-up builder run on the exact deployed behavior that needs repair.
No review appeared after deployment
The review is blocked
The review failed but the app looks correct