Memcode AI
Show HN: Memcode Launches GitHub Code Review
Last verified:
What is Memcode AI?
Memcode AI is a code review bot that analyzes pull requests against the full repository context, not just diffs. It comments only when actual problems are found, stays quiet otherwise, and can verify findings against CI test results—designed for teams that want thoughtful, low-noise code reviews.
Memcode AI pricing
Pricing model: Freemium
Memcode AI pros
- Reads entire repository context to catch cross-file dependency breaks, not just isolated diffs
- Quiet by default with strict severity filtering—only posts medium and above findings, no noise
- Evidence-driven: findings are verified against named CI tests and automatically withdrawn if disproven
- Highly configurable per repository: severity levels, tone, language, comment limits, and path ignoring
- Supports 20+ languages and checks if changes actually close the linked issues
Memcode AI cons
- Cannot write code or automatically merge—only comments; you must accept or ignore suggested fixes
- Requires GitHub App installation and per-repository access grants
- Hard ceiling of 10 comments per review; depends on CI test suite to verify findings
Frequently asked questions about Memcode AI
Does it comment on everything?
No. It's quiet by default—only medium severity and above findings are posted. Lower severity findings are recorded but kept off the PR, and the same issue is never raised twice.
How does it catch bugs the diff alone would miss?
It reads a full checkout of the repository, not just the diff, so it can find where a change alters an assumption and locate the code that depended on it.
Can it verify its own findings?
Yes. CI test results are attached to findings. If a test passes that would disprove the finding, the comment is automatically withdrawn and the thread resolved.
Can I customize how it reviews my code?
Yes. Every repository has its own settings for severity threshold, tone, language, maximum comments per review, and paths to ignore.