Install on GitHub

AI code review for GitHub

Your AI Senior Developer for GitHub.

Review every Pull Request with AI that understands your entire repository.

Senior Dev Bot analyzes code changes in the context of your existing codebase using Retrieval-Augmented Generation, then posts actionable review feedback directly on GitHub.

Works directly with GitHub Pull Requests

acme/payments-api/PR #142
Analyzing…

feat: add authentication middleware

auth.pymiddleware/session.pymodels/user.pytests/test_auth.py
def get_current_user(request):
- user = db.query(User).get(request.user_id)
+ user = db.query(User).get(request.user_id)
+ if not verify_token(request.headers['Authorization']):
+ raise Unauthorized()
return user
SSenior Dev BotPotential issue

auth.py:84

Token validation occurs after the database query. This may allow unauthorized requests to reach protected resources.

Suggested fix →[ View in GitHub ]

The problem

Code review should understand the codebase.

Traditional automated review looks at the diff in isolation. Most real defects depend on code the diff never touches.

01

Changed Lines Aren't Enough

A Pull Request can modify five lines while breaking behavior somewhere else in the repository.

02

Context Matters

A reviewer needs to understand functions, dependencies, patterns, and existing architecture.

03

Senior Developers Are a Bottleneck

Manual review takes valuable engineering time and can delay development.

Senior Dev Bot brings repository context into every review.

How it works

From Pull Request to intelligent review.

Every stage runs automatically. Developers keep working in GitHub — no new dashboard, no new workflow.

01 — Install

Add the GitHub App and select the repositories you want reviewed.

02 — Index

Repository code is chunked, embedded, and stored as searchable knowledge.

03 — Open a PR

A webhook delivers the Pull Request event the moment it is opened.

04 — Retrieve

Semantic similarity surfaces the repository code that matters for this change.

05 — Review

Findings are posted as review comments on the exact lines they concern.

Review pipeline

GitHub Pull Request
Webhook
Fetch PR Diff
Retrieve Repository Context
RAG
LLM Analysis
Actionable Findings
GitHub Review

Retrieval-augmented review

The difference is context.

Senior Dev Bot doesn't treat a Pull Request as an isolated code snippet. It retrieves relevant parts of the repository before asking the AI to review the change.

Architecture

Repository
Chunking
Gemini Embeddings
Supabase + pgvector
Vector Knowledge Base
Semantic Retrieval
Relevant Repository Context
LLM

Example

PR changes

src/auth/service.py

↓ retrieve relevant code

  • src/auth/token.py
  • src/users/models.py
  • src/middleware/auth.py
  • tests/test_auth.py

↓

AI Review

Don't just ask AI what the code does. Give it the context to understand why it matters.

Repository memory

Your repository becomes the AI's memory.

Repositories are ingested once into a vector knowledge base. Commit and base SHA awareness lets Senior Dev Bot reuse existing repository knowledge instead of reprocessing the entire codebase for every Pull Request.

Context is built once. Reviews become incremental.

Repository
Initial ingestion
Knowledge Base
PR #101→ reuse knowledge
PR #102→ reuse knowledge
PR #103→ reuse knowledge

Coverage

Catch issues before they reach production.

Bugs

Potential logic errors and incorrect behavior.

Breaking Changes

Changes that can affect existing functionality elsewhere.

Security

Suspicious or unsafe implementation patterns.

Reliability

Failure cases, edge cases, and robustness issues.

Architecture

Inconsistencies with existing repository patterns.

Code Quality

Clearer, safer, more maintainable implementations.

In GitHub

The review lives where your code lives.

Findings are posted as review comments on the Pull Request itself. Nothing is copied into another dashboard.

Senior Dev Bot reviewed this pull request
3 issues found
High

Authentication bypass possible

src/auth/service.py:84

Token validation runs after the user lookup. A request with an invalid token can still trigger a privileged query.

Medium

Missing error handling

src/middleware/auth.py:31

Session decode can raise on malformed cookies; the request will fail with a 500 instead of a 401.

Suggestion

Simplify duplicated validation logic

src/users/models.py:12

This check already exists in validators.ensure_active_user — reuse it to keep behavior consistent.

Posted automatically by Senior Dev Bot

Installation

Install once. Review automatically.

01

Install the GitHub App

02

Select your repositories

03

Open a Pull Request

Senior Dev Bot handles the review automatically.

Install Senior Dev Bot on GitHub

Stack

Built for modern engineering teams.

FastAPI
GitHub Apps
GitHub Webhooks
Groq
Gemini Embeddings
Supabase
pgvector
RAG
Python

Request path

GitHub
FastAPI
Review Pipeline
RAG + pgvector
Groq LLM
GitHub Review

Trust

Designed for your codebase.

01

GitHub remains the source of truth for your code.

02

Access is controlled entirely through GitHub App permissions.

03

Repository data is used to provide review context.

04

Reviews are posted directly to Pull Requests.

05

No manual code uploads required.

1

GitHub App

∞

Repository files understood through retrieval

24/7

Automated PR review

1

Source of truth — GitHub

Give every Pull Request a senior developer.

Install Senior Dev Bot and let AI review your code before you merge.