AI-Code-Context-Reviewer

An automated code reviewer that reviews & summarizes Pull Requests.

Tools / Libraries

• FastAPI
• Groq
• Github Webhook API Gateway + REST APIs
• Vercel (Serverless Deployment)

Python Dependency Management ( using uv )

Install uv: ( official site: uv Installation Reference↗)

curl -LsSf https://astral.sh/uv/install.sh | sh

Install dependencies (app with pyproject.toml):

cd $server
uv install

Architecture Overview

The following showcases flow diagram of how it functions:

flowchart TD
    A[GitHub Repo]
    A -- pull_request event --> B[Webhook / API Gateway]

    B --> C

    subgraph WS[Worker Subsystem]
        C[Serverless Review Worker]
        D[Context Parser Builder<br/>definitions · calls · tests]
        E[LLM Review Engine]

        C --> D
        D --> E
    end

    E --> F[GitHub Review Comments]

Loading

Current Setup / Installation

To use the reviewer with the repository:

1. Create a GitHub webhook for the repository

   • Go to Repository → Settings → Webhooks

   • Add a new webhook

   • Set Payload URL to the deployed webhook endpoint

   • Content type: application/json

   • Select individual events: Pull requests & Issue comments

2. Create a ruleset

   • Create a rules.txt file within repository following preferred guidelines for review; or refer to the existing rules in this repo for reference.

3. Open a Pull request

   • A structured summary will be automatically generated.

4. Comment /review on a Pull request

   • Inline review comments will be generated on the PR diff.