overcode

A TUI supervisor for managing multiple coding agents (Claude Code, opencode, Codex, Grok) in tmux.

Launch autonomous coding agents, monitor their progress in real-time, track costs and activity, and coordinate work across your projects—all from a single dashboard.

Why overcode?

Running multiple coding agents is powerful, but managing them gets chaotic fast. Overcode solves this by giving you:

• Unified visibility - See all agents at a glance: what they're working on, whether they need input, and how much they're costing you
• Native tmux integration - A split layout with your dashboard on top and the focused agent's live terminal below—full speed, full color, full scrollback
• Smart orchestration - An optional supervisor daemon can approve prompts and keep agents moving without constant attention
• Multi-machine monitoring - Sister integration aggregates agents from multiple machines into one view
• Session persistence - Agents run in tmux, surviving terminal disconnects. Pick up where you left off

Screenshot

Tmux split layout — Monitor all agents in the top pane. The bottom pane shows the selected agent's live terminal. Navigate with j/k and the bottom pane follows. Press Tab to switch focus between panes.

Quick Start

Try it instantly with uvx:

uvx overcode monitor

This opens the standalone dashboard. Press n to create your first agent.

For the full tmux-native experience (recommended):

pip install overcode
overcode tmux

This creates a split layout: the overcode dashboard on top, the focused agent's live terminal on the bottom. Navigate agents with j/k — the bottom pane follows automatically. Press Tab to toggle focus between panes.

Requirements: Python 3.12+, tmux, and an agent CLI — Claude Code, opencode, Codex CLI, or Grok Build

See the Getting Started Guide for a complete walkthrough.

Features

Tmux Split Layout (overcode tmux)

The recommended way to use overcode. Creates a two-pane layout in tmux:

• Top pane: Compact agent dashboard with live status
• Bottom pane: The focused agent's native terminal — no emulation, real tmux
• j/k navigates agents, bottom pane follows
• Tab toggles focus between dashboard and terminal
• Option+J/K navigates agents from the terminal pane
• PageUp/Down and mouse scroll work in the bottom pane

Real-time Dashboard

The TUI displays all agents with live status updates, showing:

• Current activity and AI-generated summaries
• Status indicators (running/waiting/stalled)
• Cost and token usage per agent
• Git repo and branch information
• Timeline showing status history

Agent Management

• Launch agents with custom prompts and permission settings
• Send instructions directly from the dashboard
• Standing orders - persistent instructions that guide agent behavior
• Agent hierarchy - parent/child delegation with follow mode and reporting
• Cost budgets - per-agent spending limits with automatic enforcement
• Sleep mode - pause agents and exclude them from stats

Supervisor Daemon

An optional Claude-powered orchestrator that:

• Monitors agents for prompts requiring approval
• Automatically handles routine confirmations
• Follows per-agent standing orders
• Tracks interventions and steering decisions

Wrappers

Run agents in custom environments — containers, VMs, or any setup your project needs:

• Devcontainer wrapper - Launch agents inside Docker containers with a single flag
• Auto-install - Bundled wrappers install themselves on first use
• Customisable - Write your own wrapper script or modify the bundled ones
• Set per-agent (--wrapper devcontainer) or as default in config

See the Wrappers Guide for setup and customisation.

Agent Backends

Overcode is not Claude-Code-only. It supports four agent CLIs today: Claude Code, opencode, Codex, and Grok. Launch a non-default one with overcode launch -n my-agent --backend opencode (or codex / grok) and it appears in the same dashboard with live hook-grade status, previews, AI summaries, send-instruction, restart, kill, resume, fork, and token/cost/context columns. Backends declare their capabilities, so the Claude-only subsystems — skills, the sandbox badge, the subscription-usage widget, agent teams — are hidden rather than shown as misleading zeros. A backend badge column appears only when your fleet actually mixes backends, so a Claude-only setup looks exactly as it did. See the Backends Guide for the full support matrix, flag mapping, and current limitations.

Sister Integration

Aggregate agents from multiple machines into one dashboard:

• Configure sister machines in ~/.overcode/config.yaml
• Remote agents appear alongside local ones
• In tmux split mode, selecting a sister agent auto-zooms the dashboard with a preview pane

Analytics & Export

• Web dashboard - mobile-friendly monitoring from any device
• Historical analytics - browse session history with charts
• Parquet export - analyze data in Jupyter notebooks
• Presence tracking - correlate activity with your availability

TUI Controls

See the TUI Guide for all keyboard shortcuts.

Documentation

• Getting Started - Installation and first steps
• CLI Reference - All commands and options
• TUI Guide - Keyboard shortcuts and display modes
• Configuration - Config file and environment variables
• Wrappers - Run agents in containers and custom environments
• Advanced Features - Sleep mode, handover, remote monitoring

Testing

Unit tests run on the host (make test-unit). End-to-end tests run inside a disposable Docker container so daemons, tmux servers and (mock or real) Claude processes can never leak onto your machine:

make e2e         # workflow + visual tiers (mock Claude, no credentials)
make e2e-real    # real-LLM smoke tier (needs CLAUDE_CODE_OAUTH_TOKEN)
make e2e-shell   # debug shell inside the test container

See docs/design/e2e-devcontainer-testing.md for the architecture, coverage matrix, and CI setup.

License

MIT