Optibot
Agentic security-first code review w/ clear cues & no noise
483 followers
Agentic security-first code review w/ clear cues & no noise
483 followers
Meet Optibot, the AI agent that thinks, reviews, and maintains your codebase. It lives in GitHub, reviews PRs with judgment, fixes CI issues and keeps your code clean long after it's merged. Built by Optimal AI.
Free
Launch Team / Built With
Render: Get up to $100K in startup creditsCloud infrastructure for founders who want to move fast and NOT break things
Promoted
Hey PH community! @chrismessina ty for the hunt!
Long-time lurkers, 3rd time makers :) @syedahmedz and I are YC alums and we're so excited to introduce you today to Optibot, your repo's code review agent. We're the former creators of Tara AI and have spent over a decade working with thousands of engineering teams.
With vibe coding, we realized that code maintenance and security would only get harder for engineering teams. After trying every code review agent and tool out there, we were surprised by how much noise they added and how it felt like they didn't have code context or reasoning.
And so, Optimal AI was born, with a mission to help maintain codebases securely. We released our first agent, Optibot, in private beta to 2,900 engineers at fortune 500 companies and startups. Today, we're excited to introduce you to Optibot's public beta after (much!) testing. We spent 8 months nailing reasoning and giving it the ability to think before responding, since the hardest problem with having truly capable agents, is context and judgment. In the video, we share how we battle tested Copilot reviews vs Optibot, and {spoiler alert} how it found 4 issues and a breaking error vs Copilot finding 2 issues in a PR.
Here's how it works:
-> Install on Github
And that's it. Once added to your repos, Optibot spends time understanding your codebase. You can speak to it directly in Github comments and ask it questions. It will summarize your PRs and review them. It's already SOC 2 verified and in use at public companies today (ty to our customers!).
PH community, we're going to be live all hours of today on rounds of coffee! Pls ask us any questions your heart fancies, and we'll respond. Oh and we're also announcing our $2.25M pre-seed raise today to help serve our customers and continue to build more agents with context.
May the bots live on to serve us (and not the other way round).
- Iba and Syed
I love that it brings real context and judgment to PR reviews - that’s a big leap beyond the typical automated tools. How customizable is it when it comes to tailoring review rules to fit specific code standards or team needs?
@evgenii_zaitsev1 Thanks! That’s exactly the gap we’re trying to bridge, moving from automated rules to thoughtful judgment.
Optibot is built to be highly adaptive to its environment. It picks up on team specific coding practices by reading existing docs, code patterns, and historical decisions, so it naturally aligns with your org’s standards without needing heavy manual setup.
But it goes even further. It also observes how team members communicate. Whether your team leans more direct or collaborative in tone, Optibot mirrors that, making it feel like a real colleague, not a rigid bot. The goal is for it to blend in seamlessly, like a puzzle piece that just fits.
Customization continues to improve as we build, and we’d love to hear how you'd like to see it evolve too!
Ty@evgenii_zaitsev1 ,
Great question! Customization is a key feature. Here’s how it works:
For Your Code Standards: Simply add a guidelines document (like `CONTRIBUTING.md`) to your repo. Optibot will learn and enforce your specific rules from that file.
For Bot Behavior: Use the `.optibot` config file to toggle features like auto-reviews, PR summaries, and auto-approval on or off.
It's designed to be very flexible to fit your team's specific needs.
@evgenii_zaitsev1 great question!
Yes! Optibot is fully customizable.
You can define one or multiple guideline files in your configuration, Optibot will then automatically follow and enforce them during pull request reviews.
This allows you to make sure all your code review process matches your team's coding standards, best practices, and project-specific rules.
Awesome product. In the new world of vibe and agentic coding, this is more important than ever.
Sold! Where do I put my credit card? 🤑
@n_wentz Hey nick!
You can sign up here: https://getoptimal.ai/
@n_wentz you're feeding a family of optibots with your credit card Nick.
Congrats on the launch! it's been really exciting to work with you during the development of such a compelling tool!
@katft Kat - it’s been amazing having you folks as our earliest customers and supporters, thank you! We’ve loved working with the Finetune and Prometric teams in refining insights. And your feedback has been vital to shaping how we think about dev productivity and speeding up projects. Thanks again!
This is a cool idea - but I'm a little concerned about the ToS/Privacy policy. It says that this grants Optibot an unrestricted license to my source code/"submission"? Can you confirm or deny that? What is the storage or usage of the codebase used in question?
@jordan_dominguez Hi Jordan. We do not store any of your code or train Optibot on any of your code submissions. We use these permissions to be able to leave a code review within github, access your customization file called .optibot (if you were to set one up, we only read this and not store it) and lastly it helps us access your code diff in a PR. We're also SOC2 certified and used by public companies.
@syedahmedz Alright! I'll give it a shot. :)
@syedahmedz @jordan_dominguez Thanks Jordan! And I'd be happy to take you through a live 15 min demo
So excited to see this out in the wild! As an engineer on the team, one of my favorite things is seeing Optibot act like a true teammate, and I just had to share this interaction from last week.
I addressed a review from it and casually told Optibot it was done. It basically replied with, "Are you sure?"
It proceeded to point out that I had only partially solved the problem. The bot's ability to understand the nuance between a technically "working" fix and a good fix (i.e., actually showing an error to the user) is incredible.
It's moments like this—where the bot catches something a human reviewer might easily miss in a hurry—that make me so proud of this product. It's like having a tireless pair programmer who helps enforce our coding standards 24/7.
Hope you all enjoy using it and happy to answer any questions !
As a small ai company, looking to stay small, its solutions like Optibot that
peak my interest. I'd love to ensure my engineers can focus on building and expanding our product, while important, but menial tasks are accomplished by agents.
That said, our customers are going to love this too, as most of them vibe code their way to success! Looking forward to trying this out and recommending to our friends.
@michaelthoy Michael, thank you for the thoughtful response! A vibe wave at you right back.