JACoB – Open-Source AI Coding Agent for Real-World Productivity
Hi, we're Kevin, Chris, and Brian, the creators of JACoB (Just Another Coding Bot), an open-source AI coding agent designed to streamline the development process. As veteran developers, we know that while AI isn't perfect, it can be a valuable tool for tasks like code reviews, converting designs to React components, and writing high-quality GitHub issues. We’ve been building JACoB over the past year to assist with our real work, and today we’re open sourcing it under an Apache 2.0 license. Our goal is to create the first fully open-source tool that matches the quality and polish of many…
In plain words
JACoB is an open-source AI coding agent built by three veteran developers to help with routine development tasks. It performs code reviews, converts design mockups to React components, generates boilerplate code, and writes GitHub issues. The tool is designed for practical developer productivity rather than replacing human judgment, focusing on specific time-consuming tasks where AI performs reliably. JACoB is available under an Apache 2.0 license on GitHub.
written from the facts on this page · September 2026
From the sources
In the maker’s words, at launch
Hi, we're Kevin, Chris, and Brian, the creators of JACoB (Just Another Coding Bot), an open-source AI coding agent designed to streamline the development process. As veteran developers, we know that while AI isn't perfect, it can be a valuable tool for tasks like code reviews, converting designs to React components, and writing high-quality GitHub issues. We’ve been building JACoB over the past year to assist with our real work, and today we’re open sourcing it under an Apache 2.0 license. Our goal is to create the first fully open-source tool that matches the quality and polish of many closed-source alternatives. JACoB focuses on specific, time-consuming tasks that AI handles well, like generating code from Figma designs, fixing simple bugs, and writing boilerplate code like API endpoints. You can even converse with JACoB to refine existing GitHub issues or help write new ones. It integrates seamlessly into your workflow, from GitHub issues to pull requests, ensuring all code passes tests before finalizing. JACoB is available on GitHub today, where you can customize it to your needs, such as running with local models. We're also developing a hosted version - sign up for the waitlist at https://jacb.ai. Currently, JACoB works best with JavaScript or TypeScript projects, especially with frameworks like Next.js. To evaluate JACoB’s real-world abilities, we created our own benchmark arena since none existed for design-to-code tools. JACoB outperformed eight popular tools and human-written code, scoring an ELO of 1183, over 100 points higher than the nearest competitor. More details on how we achieved this are available in the technical whitepaper found in the repo. AI is transforming software development, and we believe developers should lead this change, not large corporations. JACoB is our contribution to that future, respecting your privacy and security while remaining transparent and open. We'd love your feedback. Visit our GitHub repo to try it out and find more demo videos and a link to our technical whitepaper. https://github.com/jacob-ai-bot/jacob Let's build the future of AI in software development together. We’re excited to hear your thoughts and answer any questions!
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 26d ago · cactuscompute.com
