Code mode: run a coding agent on your own repository
Describe a task, get a plan, approve it, and watch the agent write code, run commands and show you every diff — on your repo, with your key, under an explicit permission model.
The More AI team · Updated June 14, 2026
A coding agent without the cloud
Code mode gives you the Claude-Code-style workflow as a desktop app: sessions are grouped by repository, each session pins a working folder, a git branch and a model. The agent explores the repo with read tools — list, read, find, grep — before it ever proposes a change.
Because it runs through your own API key, there is no per-seat agent subscription: you pay your provider for the tokens the task takes, and nothing else.
From task to plan to build
For a small change the agent just does it; for anything substantial it presents a structured plan — summary, steps, the files each step touches — and waits for your “Approve & build”. Prefer to always see a plan first? Toggle “Plan first” in the workspace bar.
During the build you see every write as a diff card with added/removed line counts, expandable to the exact lines.
Commands, tests and the permission model
The agent can run commands — install dependencies, execute your test suite — with output streaming live into the transcript. Access modes control how much it may do without asking: read-only approves every action, auto edits files freely inside the workspace but asks for risky commands, and every approval can be granted once, for the session, or always.
Repo conventions are picked up automatically: if your project ships a CLAUDE.md, AGENTS.md, .cursorrules or similar instruction file, the agent reads and follows it.
Steering, sub-agents and questions
You can type while the agent works — mid-run messages are injected into its next step, so course-corrections don’t require an abort. For wide research tasks the agent can fan out read-only sub-agents that explore in parallel and report back. And when information is genuinely missing, it asks you a structured question instead of guessing.
What Code mode is not (yet)
Honest boundaries: pull-request creation and isolated worktrees are still on the roadmap — today the agent works on your current branch and you push when you’re happy. Everything shown in the transcript is what actually ran; nothing is simulated.
Frequently asked questions
Does Code mode need a separate subscription?
No. It runs through your own API key, so you pay only for the tokens a task uses — there is no per-seat agent fee on top.
Will it change my code without approval?
For anything substantial it presents a plan and waits for “Approve & build”. Access modes and per-action approvals (once, for the session, or always) control what it may do, and risky commands always pause.
Does it follow my repository’s conventions?
Yes. If your project ships a CLAUDE.md, AGENTS.md, .cursorrules or similar instruction file, the agent reads it and follows your conventions automatically.
Which languages and projects does Code mode work with?
Any repository on your disk. The agent works with your files and runs the commands your project already uses — installing dependencies, running your test suite — so it is not tied to a particular language or framework.
Can it open pull requests or work on a separate branch?
Not yet — pull-request creation and isolated worktrees are on the roadmap. Today the agent works on your current branch and you push when you are happy; everything in the transcript is what actually ran, nothing is simulated.