Trag is an AI code review companion with a twist! It's like a linter, which can lint patterns. Trag gets as an input plain english rules and reviews them on every pull request in seconds. Move your knowledge into patterns and automate reviews with Trag.

Hey Product Hunt community! I’m Khachatur (or K for short), co-founder & CEO of Trag along with @hovo_ghevondyan1 and @movses_saponjyan . I’m sure you’ve noticed that AI in software is growing rapidly. As engineers, we feel this is the missing piece in our journey to automate everything. Now we can automate more, enabling us to focus on what truly matters. The industry has never seen such rapid growth in codebases, thanks to amazing tools like GitHub Copilot, Cursor, and others. Engineering teams of the same size now have to manage much larger code changes, reviews, and processes. Trag is an AI code review companion. One of our customers described us as a “superlinter,” which we take as a compliment! So, what does that mean? There’s a lot of tribal knowledge in engineering teams that only comes out during pair programming sessions or general code reviews. Trag automates that tribal knowledge into 100% deterministic rules and enforces them on every pull request. Think of Trag as an extra teammate who already knows where to look. What makes Trag different? Trag focuses on one simple thing: matching a written rule to the code. It doesn’t perform vague or overly general code reviews. We believe you know what’s best for your repo, and we’re here to help you automate it. We would be delighted if you’d try Trag out with your team or even on a side project. As Trag is still young and we’re early in our journey, things might feel a little off at times, but we’re always here to listen and improve based on your feedback! Here’s what we offer: 🧠 Automate your knowledge: Write plain English descriptions about your codebase and describe how you want it to look like or certain things to work. 🔎 Codebase understanding: Trag has its own custom programmatic search across the entire codebase, even spanning multiple repos. That means Trag understands your whole repo and gets the context right. 🚀 Minimal setup: Just install Trag on your repo, write the rules, and of course, open a pull request! If you want to learn more: Please visit https://usetrag.com/ Join our discord https://discord.gg/XEH5Gnhg Promo code to get 50% off for 3 months - TRAGPH Try it out yourself! We will be in the comments to respond to you folks!

@hovo_ghevondyan1 @movses_saponjyan @kh_mugh Hey guyz, great work btw. Just curious to know if the review includes the basics on coding best practices, security and prod readiness. What i am struggling with AI coders is they are able to achieve functional requirement but mess up code big time which is just to try out but never prod ready.

@hovo_ghevondyan1 @movses_saponjyan @imraju Hey! Thank you very much! That is the exact thing we want to achieve, Trag reviews the code with the rules but also does general reviews if you enable it in settings. It will review the code using multiple static analyzers for security, and depending on the language with it's standards. The main thing is that you describe the prod readiness as granular rules and whether it's human generated code or AI generated code, Trag will enforce it, and make it look like you want to!

@hovo_ghevondyan1 @movses_saponjyan @kh_mugh Thanks Khachatur. Can you point me to some documentation Trag has which helps me understand how you review prod readiness and overall best practices setting specially React / Angular / Node frameworks

@hovo_ghevondyan1 @movses_saponjyan @imraju The documentation is not fully released sorry for that, but happy to jump on a quick call and go over everything.

@hovo_ghevondyan1 @movses_saponjyan @kh_mugh Your academic toolkit is an excellent resource for students aiming to manage their grades effectively!

Congratz with the launch! The idea of defining more generic linting rules with plain English is indeed promising, will try it out on our repos.

@vadim_fedorov42 Thank you very much! Please let us know about your experience, we're here to help ;) Btw, congrats on the launch!

Here because, David from Treblle said this is a cool product to check out. :D

@savan_kharod1 ahaha, thank you very much!

@kh_mugh Definitely a cool product. Although I am not a techy, the video explains the product really well. Good job on that. :)

@savan_kharod1 Thank you very much!

I’ve been using Trag for a bit now, and honestly... it’s been pretty solid. The AI-powered code review process is a huge time-saver. Instead of spending hours manually combing through PRs, Trag does a quick pre-review and even suggests fixes, which is super helpful. I really like how you can set up customizable rules for your team’s coding standards... it’s nice when a tool adapts to you instead of the other way around. That said, it’s not perfect. The AI-generated fixes are great in theory, but sometimes... they’re a little off. I’ve had a few moments where the suggestion just didn’t fit or wasn’t as efficient as I would’ve liked. It’s helpful, but you still can’t fully rely on it replacing a human review... at least not yet. The semantic analysis, though, is where Trag really shines. It’s not just looking at surface-level stuff, it actually gets into the structure of your code and makes some smart suggestions for refactoring. That’s something most tools don’t do well... so big points for that. But... I do think the UI could use some love. It feels a little clunky when you’re trying to dig into analytics or flip between different pull requests. If they cleaned up the interface a bit, it could be a game-changer. Overall, Trag is definitely worth trying if you want to speed up your code reviews and get some AI-powered help. Just be ready to double-check the AI’s fixes... and maybe keep an eye out for some UI improvements.

Helpful featuers.idea is good enough.

Maybe it can help me optimize the code.

Helpful featuers.idea is good enough.

Maybe it can help me optimize the code.

Trag

AI Code Review companion

Do you use Trag?