Giles Grolla Alonso

CTO.new offers FREE access to top tier AI agent that is able to elaborate tasks to other agents (also free, and top tier) that code directly into your github repositories, aware of the context of each repo (it does have a setup agent!).
On the past days I've been testing it to improve on some old projects I had, and it transformed them into much better versions.
Great things it has:
-Allows review and merge into github right from the platform.

-Agent tests if the changes were successful before prompting for a merge.

-You can re-run tasks that didn't complete on the first try.

-The "tasker" AI is aware of the tasks execution, and can help re-distribute the work load.
-Surelly many more that I can't remember right now.

-it could have a mobile friendly interface (let-me vibe code from my phone easily also!)

It is free, runs from my browser, pushes directly to github, zero costs from ideation to a working product.

It is free! and you can have 2 AI agents running simultaneously, while you can freely plan next tasks with the "tasking agent".

You can provide it a empty repository and ask it to build from scratch. it can even run a deploy on it's ai coding agent, and push the resulting files to github...

The AI agents put quite a effort before timing out a task, I don't know exactly the limit, but it is fairly generous and can build some pretty impressive things in one go.

It is nativelly integrated to github (haven't tested gitlab, but I'm pretty sure it links also). It also allows you to set MCP services.

Yes, the code is used to train underlying models, and it is in their terms of service.

During launching days the system had to remove Claude from the available AIs because anthropic couldn't handle it... The limit didn't come from the platform, but from the source agent provider, so, it seems pretty limitless to me.