bloop VS chatgpt-code-plugin

https://github.com/BloopAI/bloop Is fully open source and has full text + regex search built on tantivy fyi

Thanks! Bloop really helped with the documentation! they index your code base and uses RAG locally to talk against it. It works shockingly well.

In this blog post, I’ll be comparing 3 distinct AI-first code search tools I recently came across: Cody (developed by late-stage startup, Sourcegraph), SeaGOAT (an open-source project that was trending on HN last week), and Bloop (an early-stage YC startup). I’ll be evaluating them along the dimensions of user-friendliness as well as their accuracy.

If you're confused about any of the code snippets above, you can check out bloop.ai and phind.com (along with its VSCode extension) to answer any of your questions about the repository, noting that both have free plans.

COMPANY: bloop (https://bloop.ai)

bloop (YC S21) | Rust AI Engineer | Onsite | London, UK

We're a YC + tier-1 VC backed startup, working on a AI chat assistant for software developers that can answer questions about any large codebase.

About you: you hack around with GPT on the weekend, maybe you even fine-tune some of these new open source models. But you're also a solid programmer, capable of not just prototyping a nice demo but also shipping performant code that can scale. You're probably quite good at what you do, maybe you were top of your class, or maybe you rose the ranks to your comfortable management role, but yearn for the days of writing code, redbull and deploying on a Friday.

We're a small team, looking for one individual to join us in London, in-person at least 3 days per week.

Please reach out to me at join [at] bloop [dot] ai with "HN AI Engineer" in the subject line and a note about why this is exciting and a link to your portfolio or linkedin/resume.

P.S. Our stack is Rust, so it's likely the role will go to a candidate with Rust experience but if you think you have something to offer that compensates for lack of Rust we should chat anyway!

https://github.com/BloopAI/bloop

Bro let me turn your life inside out: https://bloop.ai

Another example is Bloop, it is a code search engine built on top of tantivy and qdrant.

Bloop is a developer assistant that uses GPT-4 to answer questions about your codebase (https://github.com/BloopAI/bloop)

Shameless plug - https://github.com/kesor/chatgpt-code-plugin - also does a very good job of telling you anything you'd like about your code. Shared it on ShowHN as well here https://news.ycombinator.com/item?id=36099507

My version of loading multiple files into the context allows the UI of ChatGPT to load the files on his own, when it decides that it needs/wants to read the content of the file. And from some experimentation it would seem that each file it loads, it considers it as a separate interaction. Thus the token limit is much less of a problem, allowing to load larger pieces of code - either the whole thing, or piece by piece. https://github.com/kesor/chatgpt-code-plugin

Super interesting project! I like that you're exposing the structure of the git repo to GPT, and helping it understand and access the code in a semantically meaningful way via the function APIs.

I've been working on similar things with GPT, but in a language agnostic way. I was curious how you were parsing/lexing the files to pull out functions. I asked GPT about your repo, and it provided a helpful summary of that process:

  $ git clone https://github.com/kesor/chatgpt-code-plugin.git

You can take it for a spin yourself https://github.com/kesor/chatgpt-code-plugin