Finally, a Fully Autonomous AI Coding Agent 🚀 .

TLDR:

yes, Agen is already building itself at this point 🤩
runs in the cloud, doesn’t need local installation
automatically fixes your pipelines
delivers a fully working code!
as a ProductHunt user, you can try Agen for free, and get $20 Credits on sign up

AI already changed a lot the way we code, but it didn’t really change the way we work and think about the code. For the biggest part, we have the same processes and workflows, the biggest difference being the fact that an AI model writes the code for you.

We built Agen as the coding agent for the future, with the goal to help the teams build faster, better, and importantly, secure code.

Here are the principles that are driving us, and which are the foundations of what we’ve built so far and the many things we plan to build in the future, keep reading to see what this means for you and your company.

🚀 Fully Autonomous AI Coding Agents - our goal is to make sure that the agents are being able to do the work independently, with minimum human guidance. We noticed that the biggest bottleneck are the humans right now, because the agents still need quite a lot of guidance, starting from the environment they run it, and ending with the work they do. Continue reading to see what we’re doing to fix this.

☁️ Cloud-First: this is the first step to have autonomous agents, they need an environment to run in, and right now, if you close your machine, your agent stops working. By making the agents cloud-first, we make sure that they always have a place to run in, and that they’re not blocked by humans when they need to do the work. For you and your company this means that the agents will do meaningful work 24/7 and help you move faster.

🧑‍🔧 Self-healing pipelines - this is the next step, and something that the agents are blocked by when they make changes to the codebase pretty often. Usually, a developer has to tell the agent that the pipeline failed, why it failed, and ask them to fix it, or they use the MCP (if the version control system they use has one), but it’s still happening manually, on user request. Our agents are monitoring the failed pipelines and fixing them automatically, with zero human involvement. This results in a huge amount of time saved - time that can be spent on more meaningful work.

📆 Scheduled Agents - they're running on your schedule, and they're doing autonomous work in the cloud, without human supervision - all you have to do is check and merge the code.

🪾 Multi-repository support - one of the biggest issues with the autonomous coding agents is that they're tied to a specific repository, and they can't make changes to multiple repositories for a task, usually you have to start new agents if your task involves making changes to multiple repositories. We built Agen to work flawlessly with multiple repositories, it will automatically choose the right repository to change, and it's able to make changes and open MRs in multiple repositories within the same session.

💻 Multi-session environment - one more blocker for real autonomous work, and the only way to have multiple sessions with the current agent is to create multiple copies of your repositories, and run them in different terminals, but that’s time consuming, and not efficient at all, especially when it comes to cleaning them up, creating them again, etc., All our Agents are doing the work within a session, which gets an isolated sandbox, it’s own branch, MR, etc. For you, as a developer, it’s important to be able to run multiple tasks at the same time, the work is not linear, and being able to run multiple sessions is a huge productivity boost.

📱 Mobile-friendly - you know that message you’re getting from your boss about a bug when you don’t have the computer with you? It’s happened to all of us, and it’s stressful 😣 . We optimized our web app and made it mobile friendly so that you can make changes wherever you are, and whenever you have to.

🧑‍💻👩‍💻 Built for meaningful collaboration, with per-repository access, and granular permissions, that resemble how the teams usually work. Each person can be given access to specific repositories, and they can have different roles for each repository. This keeps your codebase confidential and secure.

✅ See your pipelines & merge straight from the session, without having to switch tabs. You’ll see a green indicator when all the pipelines pass, so that you know that the work is completely finished. The next step is to actually merge the code, and you can do it without leaving Agen, straight from the session page.

🔐 Security - the Agents are making changes on your codebases in new, secured and protected sandboxes for every session. The sandboxes are destroyed once the Agent has finished the work, together with your code and any traces of it.

⏭️ What's coming next: integrations, more autonomy

Your feedback means a lot - please leave a comment and tell us what is the biggest blocker you’re having when using AI tools for coding, and what would you like to improve in your workflows.

Thank you for reading so far!

Daniel, Co-Founder 👋

Finally, a Fully Autonomous AI Coding Agent 🚀 .

TLDR:

yes, Agen is already building itself at this point 🤩
runs in the cloud, doesn’t need local installation
automatically fixes your pipelines
delivers a fully working code!
as a ProductHunt user, you can try Agen for free, and get $20 Credits on sign up

📆 Scheduled Agents - they're running on your schedule, and they're doing autonomous work in the cloud, without human supervision - all you have to do is check and merge the code.