Chorus

The deciding trait is that the model which wrote the code is not the model that clears it. Two to four rival CLIs read the same diff and the run only goes green when they agree, so a confident mistake from one vendor has to survive the others before it reaches you.

The cost is patience: four opinions take four times as long to produce and someone still has to read the disagreement. Pick it when a bad merge is expensive and the review queue is the bottleneck. Pick a single hosted reviewer when you want a comment on the pull request and nothing more.

The workaround is the problem. One of the CLIs it drives blocks protocol tools in its non-interactive mode, and the stated remedy is a flag whose name says what it does. A review tool that requires approvals and the protections around them to be turned off in order to run automatically has inverted its own purpose.

The row records this plainly: a documented hazard is still a hazard, and the flag will be pasted into a pipeline by someone who did not read the sentence around it. What it does right is naming the limitation instead of leaving it to be discovered in a failing job.

1. Consensus is a defensible proxy. Two models trained on overlapping corpora may agree on a wrong answer, so agreement bounds independent error only to the extent that the reviewers are genuinely independent, and no analysis of that independence is offered. 2. Disagreement is treated as a signal, which is the sounder half of the design.

2. The measurement that would settle this is available and unperformed: a set of pull requests with known defects, reviewed by one model and by the panel, reporting how many defects each caught. Until that exists, the claim is that several opinions beat one, which is plausible and not evidence.

The economics are borrowed. Running automated review through consumer subscriptions converts a metered cost into a flat one, and every vendor whose plan is being used this way has both the incentive and the terms to stop it. That is not a moat, it is an arbitrage, and arbitrages close.

There is also no revenue here, no hosted tier and no entity, so nothing survives the day one provider changes a rate limit. Likely path: the value proposition halves quietly and the project keeps working for whoever still holds the right plans. Position: adopt it for the review quality, and budget for the day it needs real API keys.

Sixty developers, one npm command, no licence line: the finance answer takes ten seconds. The operational answer takes longer, because a globally installed package on sixty machines is sixty versions unless someone owns the pinning, and nothing here reports centrally on what it found or what it cost.

It does run unattended, which puts it ahead of most of this shelf. No single sign-on, no directory sync, no audit export. Approved with conditions: pinned through our own package policy, run in CI rather than on desks, and reporting into the same dashboard as every other gate.

Registering itself as an MCP server with everything already installed is the correct move: nine tools appear inside the assistants I use, so a review run is something I trigger from where I am rather than from a separate window. That is the difference between a tool and a destination.

The daemon is local and the run history stays with it, so nothing about my diffs leaves the machine and there is no service to lose access to. Apache-2.0 keeps the fork available. Grudging respect: this was built by somebody who has been burned by a vendor before.