Run Claude Code from your phone

Treena now lets users run Claude Code from a phone by providing a real cloud workspace with a filesystem, shell, git, and cloned repo, streaming the agent's session, tool calls, and diffs to a mobile screen. Users sign in with GitHub, connect their own Claude account, and can start a session to edit files, run tests, and push commits without a laptop. The service is designed for steering agentic coding loops on the go, with plan mode for one-tap approval and agent mode for autonomous execution.

Run Claude Code from your phone How to use Claude Code away from your desk — start a session on your repo from a phone, steer an agentic coding loop in a real cloud workspace, and push a reviewed commit. Treena Claude Code is a terminal-native coding agent. Terminals live on computers. So the assumption goes: no laptop, no Claude Code. That assumption is wrong. The terminal doesn't need to be in front of you — it needs to be reachable. The setup Treena gives every user a real cloud workspace: a dedicated machine with a filesystem, a shell, git, and your repo cloned from GitHub. Claude Code runs there, in an actual terminal, exactly the way it runs on your laptop. Your phone is the interface. The agent's session, its tool calls, its diffs — all of it streams to a screen that fits in one hand. - Sign in with GitHub and pick a repo - Connect your own Claude account — bring-your-own-key, no model tax - Start a session: describe the task, or plan it first - Watch edits, test runs, and command output as they happen - Review the diff and push when it's right Why this beats a chat app You can paste code into any chatbot from a phone. That was never the hard part. The hard part is everything around the suggestion: Context. Claude Code reads the actual repo, not the fragment you pasted. Tools. It edits files, runs the test suite, and reacts to real errors. Consequences. The output is a commit on a branch, not a snippet in a transcript. A chat transcript is advice. An agent in a workspace is work. Steering matters more than typing On a phone you type less, so the loop has to respect that. Plan mode drafts the approach before anything is written — approve it in one tap. Agent mode runs until done or blocked. Either way, your job is judgment: read the plan, read the diff, decide. That's the honest shape of agentic development everywhere, on every screen. The phone just makes it obvious. When this is genuinely useful - The test suite broke while you're at lunch - An idea arrives on the train and you want it drafted before it fades - A teammate needs a one-line fix reviewed and shipped, and your laptop is at home - You want a long refactor started now so it's ready for review at your desk None of these deserve a "when I'm back at the computer." They deserve fifteen minutes of steering from wherever you are. Join the waitlist / waitlist and put Claude Code in your pocket.