Hey Product Hunt,

I built OpenMergeLens because I wanted automated AI review on the PRs waiting

for me, without asking every repository to install another bot or commit

workflow YAML.

OpenMergeLens is an MIT-licensed CLI that runs on your machine. It uses the

GitHub CLI accounts and repositories you explicitly select, then works with a

reviewer you already use: Codex, Claude Code, or a compatible MCP-enabled

command. Run it on demand or schedule it during setup. There is no

OpenMergeLens GitHub App, webhook, tunnel, or server.

The safety path I wanted for my own reviews:

- `openmergelens --dry-run` performs the inspection with zero GitHub mutations

- live runs validate inline anchors against the real diff and move anything

unanchored into the summary

- constrained read-only inspection is scoped to the selected PR and

repository; incomplete coverage is retried once, then fails closed without

posting

- reviews are non-binding comments; OpenMergeLens never approves a PR or

requests changes automatically

Each repository gets an editable review prompt. Explicit corrections stay

scoped to one reviewer identity and repository. Polls with results also create

private local HTML reports you can open from the desktop notification or with

`openmergelens report`; the reports contain outcomes and PR metadata, not

diffs, findings, review bodies, credentials, or tokens.

OpenMergeLens itself has no hosted service, account system, analytics, or

telemetry. PR content is still processed by the AI provider you choose and is

subject to that provider’s privacy, retention, and billing terms.

Try the safe path:

npm install -g openmergelens

openmergelens init

openmergelens --dry-run

I’d love blunt feedback on the trust model. What would make you comfortable or

uncomfortable letting a local scheduler review on your behalf?