Show HN: AI that deeply understands you codebase | Hacker News Reader

Show HN: AI that deeply understands you codebase

5 pointsaditya10817 comments

Hi everyone!

We just launched Depth AI - a tool that helps you onboard to large and messy codebases. Unlike most dev tools that help in codegen and building smaller apps, this one mainly aims at understanding large repos better - so we have focussed a lot of code search quality.

We also launched the first version on product hunt https://www.producthunt.com/posts/depth-ai. Do check us out.

Would love to hear feedback here and discuss more how our approach to code search is different.

View on HN

URL: https://www.trydepth.ai/

Adding so it doesn't take 3 clicks to get there.

Can you speak towards code privacy and security as well as pricing? All your marketing collateral is light on these details.

Yeah the PH launch was just aimed at getting some bottom up users, we ideally want to sell top down. As far as code privacy and security is concerned, we deploy depth ai in your private vpc, so your code never leaves your cloud.

We just launched, so we are looking for design partners and / or paid pilots (pricing would depend on number of users and we can discuss it more case by case).

Lmk if that answers your questions!

Can you speak more to the technical details on how you develop the codebase context construction?

How does your approach differ from what the current IDEs for AI coding already do such as the use of tree-sitter (Codebuff) or vectorizing the repo (Cursor)?

Also, what LLM model is being used for the conversations?

Yeah thats the main differentiation - our backgrounds are in search, and we think its pretty hard to use traditional chunk -> embed -> retrieve- > rerank for structured data (they work well for unstructured data).

So we use a different approach where we create the code graph and apply classic ML (like clustering) to find abstracted patterns in the graph. We have AI agents that traverse these abstractions to create knowledge which we search on.

This enables better quality responses for more complex questions about generic higher level concepts.

We think of it in the following way: Cursor etc will be extensively used to write code. Our tool comes into play when you don't know exactly where to start, which is why we focus on KPIs like new eng onboarding times, reduced eng hours on KT, etc - overall developer efficiency. Lmk if it makes sense!

Definitely, thanks for the details. What do you mean by design partners?

Hi, co-founder here (OP went to sleep lol). We are looking for companies with 50-100 devs where we can deploy our product and develop the product based on their feedback.

what is the different with aider project, I thought they are doing the same work with you.

Reply on news.ycombinator.com