AI PR Reviewer
An LLM-driven pull-request reviewer as a GitHub Action — inline comments, severity gating, no infra.
AI PR Reviewer is a composite GitHub Action that runs a real LLM-driven code review on every pull request. It posts inline comments and GitHub suggestion blocks, marks previous bot reviews as outdated so your PR page stays readable, gates the CI check by severity (critical / warning / info), and applies a `reviewed` label when done. It's stdlib-only Python — no Docker image to pull, no Node modules to install, no infrastructure to maintain. Add one `uses:` line and your PRs get a second reviewer that never sleeps.
1
line of YAML to enable
0
infra required — pure GitHub Action
3
severity levels with configurable gating
MIT
licensed, end to end
Why it exists
The bits that make it worth adopting.
Inline comments + GitHub suggestions
The reviewer posts comments right where the code changed, with GitHub's native suggestion syntax when it can propose a concrete fix. Feels like a human reviewer using the same tools.
Severity-based gating
Every finding is `critical`, `warning`, or `info`. Configure which severities gate the CI check. Ship faster on info-only reviews, block on critical.
Old-review outdating
Push a new commit and the reviewer marks previous bot reviews as outdated. Your PR page stays readable — only the current review is expanded.
Zero infra to run
No Docker image to pull, no Node modules to install, no server to host. GitHub runs the action, the action calls your LLM provider, the reviewer posts back to GitHub.
Provider-agnostic (roadmap)
Anthropic today. OpenAI and Gemini on the roadmap. Configure via `provider:` input. Bring your own API key, keep control of your data.
MIT + fully forkable
The whole action is MIT-licensed. Fork it, wire it to a private model, extend the severity model — the code and licence say yes.
Install
Get started in seconds.
Pick the channel that matches your stack. Every path lands you in the same working state.
uses: DailybotHQ/ai-pr-reviewer@v1In practice
What teams actually do with it.
01
Fast catch-up on medium-sized PRs
Reviewers still do the thoughtful pass, but they skip the class of issues that don't need a human eye — obvious typos, missing null checks, style drift. The bot flags those inline before the human even opens the tab.
02
Standards enforcement without a Slack fight
Configure severity thresholds and let the bot gate the CI. Teams stop arguing about style over Slack — the check either fails on critical, or it doesn't.
03
New-contributor friendliness
First-time contributors get instant, non-judgmental feedback. Nice tone, actionable suggestions, no waiting hours for a maintainer.
04
Long weekends without silent PRs
PRs opened Friday afternoon get a review before Monday morning. The team can still merge safe changes, and the human review queue starts smaller Monday.
At a glance
The short list.
Inline comments + GitHub suggestion blocks
Severity-based gating (critical / warning / info)
Auto-collapses previous bot reviews
Anthropic today, OpenAI + Gemini on the roadmap
FAQ
The questions we hear the most.
Which LLM providers are supported?
Anthropic today. OpenAI and Google Gemini are on the roadmap. The provider is a config input — you bring your own key.
Where does the code get sent?
To whichever LLM provider you configure — the action never proxies through Dailybot. Your API key, your data, your choice of provider.
Does it replace human review?
No. It handles the class of feedback that doesn't need a human eye, so humans focus on architecture, correctness, and taste. Think of it as a really thorough first pass.
What it is
AI PR Reviewer is a composite GitHub Action that adds an LLM-driven reviewer to every pull request. It posts inline comments, gates the CI check by severity, and applies a reviewed label — all with a single uses: line in your workflow. Stdlib-only Python underneath, so there’s no infrastructure to maintain.
Why open source
Because “which reviewer is looking at my code?” should have an answer you can inspect. MIT-licensed, developed in the open, no black-box behavior. Fork it, wire it to a private model, extend the severity contract — the code and licence say yes.
How it fits the ecosystem
Dailybot uses this action to review its own PRs (including the ones AI agents open). It’s the kind of small, focused tool that fits the same pattern as the rest of DailybotHQ: MIT, minimal deps, and no infra you have to run.
Ready to try it?
Open source, MIT-licensed, and shipping in production at Dailybot every day. Fork it, wire it in, contribute back.
Other Dailybot open source
Dailybot CLI
The command-line client for the whole Dailybot API.
Dailybot Agent Skill
The official adapter between your AI coding agent and Dailybot's public API.
Deep Work Plan
The gated-plan methodology any AI coding agent can execute.
Deep Work Plan website
The open-source marketing site and documentation for the Deep Work Plan methodology.