Thread by @mattzcarey

Introducing Code Review GPT: Your personal code reviewer powered by LLMs⚡️

Improve code quality and catch bugs before you break production 🚀

How does it work?

🧵 👇

Spoiler: we used @LangChainAI from @hwchase17 and it's incredible.

Code Review GPT runs in CI/CD pipeline.

Provides super detailed and helpful suggestions on your code 🤝

It gets all the changed files using this lovely git command:

`git diff --name-only --diff-filter=AMT ${baseSha} ${githubSha}`

And it builds a fancy prompt to tell the LLM exactly what to do 🤩

Code Review GPT uses @LangChainAI to make requests to the LLM.

In this case, @OpenAI GPT-4 works pretty well and now everyone can use it >>> openai.com/blog/gpt-4-api…

Using the brand new @LangChainAI Smith tool, you get this awesome dashboard 💫

Keep track of your costs and responses from Code Review GPT super easily.

Code Review GPT then rates the responses and posts the most severe problems on your PR 👌🏼

Watch the demo:

Roadmap is full 🌝

👀 Prompt performance testing tool 🤯

👀 Vector db to give the LLM memory for even better suggestions

What to pick? @neondata @supabase @awscloud Aurora Serverless... @trychroma @SingleStoreDB @momentohq @pinecone

Let me know below 😬🤟

From the team building @quivr_brain @_StanGirard

You can see Code Review GPT in action there:

@quivr_brain @_StanGirard Repo link: github.com/mattzcarey/cod…