Merger verdicts
Merge, reject, or escalate, weighing CI, findings, blast radius, and history. Risky changes go to a human.
Merge Automation
Automate the confirmations. Measure the result.
Merger returns a verdict on every open pull request. Your rules decide what clears without a human, and Insights shows whether it works.
Decide, automate, measure
Merger makes the call, your rules bound it, and the dashboards tell you when to widen them.
- GateDeterministic rules per repository
- DecideMerge, reject, or escalate, with rationale
- AutomateResolve, approve, and merge what qualifies
- MeasureVerdicts, acceptance, bugs caught, time to merge
Rules that set the bar
Hard gates plus written guidance per repository. Merge automatically or just label.
Auto-resolve and auto-approve
Resolve addressed threads and approve PRs that meet your conditions.
When and where review runs
Triggers, checks, comment limits, and skipped paths, per repository.
Merger Agent Stats
Evaluated PRs, ready-to-merge rate, and the share merged without a human.
Agent evaluation
Accepted and rejected suggestions per agent.
Engineering impact
Bugs caught versus reported, and time to merge against a pre-Baz baseline.
Attention where the risk is
Every PR treated alikeMerge Automation
Every change waits for someone to re-check CI and commentsChanges that meet your rules merge or clear on their own
Addressed threads linger until someone closes themAddressed findings resolve automatically
High-risk changes get the same skim as typo fixesUncertain or high-impact changes escalate to a human
"Is the automation working?" is a guessVerdicts, acceptance, bugs caught, and time to merge, measured
Questions teams ask first
Does Merger merge without a human?
Only where a rule allows it. Uncertain changes always escalate.
Do the dashboards need setup?
No. They compute from review activity.
Can we start without automation?
Yes. Start manual, then widen the rules as the numbers earn it.
Put merges on rules you can measure
Connect a repository and see Merger's verdicts and your numbers on real PRs, or ask us how teams like yours set the rules.