vibepanel CLI

A web console for developers running several agents across projects: sessions survive crashes and replies come from a phone.

  • 57 ★ on the repository
  • Last commit
CLI

Broad access

Why this level

  • The panel gives web access to your machine terminals, protected only by a password or passkey.
  • Chat replies approve agent commands, and hooks edit the settings of Claude Code and other agents.
All reasons and checks

jiangmuran/vibepanel

Install

Manual install

curl -fsSL https://raw.githubusercontent.com/jiangmuran/vibepanel/main/install.sh | sh

The standard interactive installer for Linux and macOS, the first README command.

Access and checks

How we review

Broad accessThe tool can change data, run commands or handle secrets.

  • The panel gives web access to your machine terminals, protected only by a password or passkey.
  • Chat replies approve agent commands, and hooks edit the settings of Claude Code and other agents.

This is third-party code. Review the repository files before installing.

What it does

vibepanel is a working terminal for people who lean on agents hard. Terminals live in a separate system service, so a session survives a front-end crash. Upgrading or restarting the panel touches no running session. The interface opens as a web page, so you can reach a development machine from any device, and there is a separate phone version. The whole model is projects and sessions: you see at a glance which session finished, is working or waits on you. A file manager and notebook sit on the right and a scratch terminal along the bottom. A hook for Claude Code, Codex, Kimi Code, zcode or opencode reports that an agent is really done. A permission request arrives as a card in Telegram, Feishu or WeChat and is answered right in the chat. The same binary doubles as the admin CLI.

Who it is for. Developers with several terminal agents at once and a machine that should run around the clock and be reachable from any device.

Good fit when

  • Five agent sessions run across projects and you lose track of which one waits for you.
  • An agent stopped on a question while you left the house and you must answer from a phone.
  • The work machine is remote and terminals must survive panel upgrades and restarts.
  • A screen or monitor needs a summary of token spend and session state.

Not a fit when

  • You need commercial use: the licence is PolyForm Noncommercial and business use needs a separate one.
  • A single terminal and one agent are enough.
  • Policy forbids opening web access to a development machine.

Example request

Show the state of all agent sessions by project and list the ones waiting for my reply.

Limitations

The licence is PolyForm Noncommercial 1.0.0 with attribution terms: personal and noncommercial use is free, commercial use needs a separate licence. Releases up to and including 1.20.1 shipped under MIT. The project is early, as the author notes. Installation runs the repository `install.sh` script through `curl`. The panel accepts network connections, so protect login with a password or passkey. Telegram, Feishu and WeChat integrations are off by default and are enabled by you after consent.

How to disable. Stop and remove the service with `vibepanel service uninstall`. Remove agent hooks with the `hook` subcommand or by hand; the backup made before installation helps roll them back.

FAQ

How do I install vibepanel?

On Linux and macOS there is an interactive installer: `curl -fsSL https://raw.githubusercontent.com/jiangmuran/vibepanel/main/install.sh | sh`. Details, Docker and building from source are in `docs/install.md`.

How do I manage the panel after install?

With the `vibepanel service` command and the `status`, `start`, `stop`, `restart`, `logs`, `token`, `upgrade` or `uninstall` arguments.

How do I know an agent has finished?

One click in the panel installs a hook into Claude Code, Codex, Kimi Code, zcode or opencode. The panel shows what it will write and backs up the file first.

How do I reply to an agent from a phone?

Pair the panel with a Telegram bot, a Feishu app or personal WeChat. A card with the session number arrives in the chat, and a reply like `3: add a test` lands in that session.

Can I drive the panel from a program?

Yes. The panel exposes an HTTP API with a token from settings. The `project`, `session`, `hook` and other subcommands work through the same API.

Discussion

No questions or reviews yet.

CLI tools

A local control panel for coding agents: it runs Claude Code and Codex, keeps a task queue and merges verified work into main.

CLIBroad access119Repository stars

A tmux-based TUI that runs Claude Code, Codex and other agents in one list: status, quick prompts, worktrees and diff review

CLIBroad access691Repository stars

Mjolnir runs Claude Code, Codex and other agent sessions in a terminal, in containers and over SSH, and keeps them running.

CLIBroad access115Repository stars

A pixel-art office in the browser: each Claude Code or Codex session sits at a desk, types and raises a hand when it needs you.

PluginExtended access60Repository stars

A terminal workspace where several CLI agents like Codex, Claude and Gemini exchange messages and work on one task under human oversight

CLIExtended access3.6KRepository stars

Viden is a terminal cockpit for a coding agent: chat with a model, approve edits and supervise Codex and Claude in one window.

CLIBroad access86Repository stars

README in short

The README explains how the panel works, shows interface screenshots and covers chat sign-in. Installation, service commands, licence and the development check list follow.

README badge

Report a problem