I'm an academic who also codes, so my daily usage is messier than a pure developer's: prototyping code, rearranging project files, revising parts of an academic paper, and occasionally asking it to diagnose something weird on my laptop. Freebuff handles that mixed workflow well. I've used the web app, the CLI, and the desktop app, and for real work I'd rank them desktop first, CLI second, web last. The desktop app working directly with my file system is a big deal for me. Even with my country sitting in the restricted access mode, it has quietly replaced several paid AI subscriptions I used to maintain.
The sidebar in the screenshot shows "Freebucks 75", while the homepage says unlimited usage. What does a Freebuck count, and what happens when it hits zero?
I pay for Claude Code daily so the "free" pitch immediately makes me skeptical, but the multi-model picker is the part that caught my eye — being able to switch off whatever the default free model is matters a lot when a model starts going sideways mid-task. Curious how the open models hold up on a large existing codebase rather than a fresh project — that's where most free agents fall apart for me. Congrats on the launch!
Running it from the terminal against a project I already have open is what would get me to try this, since every other free coding agent wants me to move my work into their editor first. Does the terminal mode see my existing repo the same way the browser one does?