Tmuxlet

Tmuxlet is an open-source terminal agent harness and orchestration tool developed by CodefiLabs. It provides an isolation, supervision, and execution layer that coordinates interactive CLI coding agents inside dedicated tmux sessions. In daily use, Tmuxlet enables developers and automated scripts to dispatch tasks to CLI-based coding agents (such as Claude Code and Agy) without stalling the active shell. It manages process lifecycles inside isolated tmux sessions identified by distinct hashes, routes prompts via command arguments or stdin pipes using the standard `-` sentinel, and handles session continuation using flags like `--continue` or `--conversation <id>`. What distinguishes Tmuxlet from typical wrapper scripts is its built-in infrastructure protection. It implements a live-session semaphore to enforce machine-wide concurrency caps (defaulting to 24), a session reaper (`tmuxlet reap`) to clean up hung or zombie agent sessions, paced input broadcasting (`tmuxlet send-all`) to mitigate thundering-herd issues on mass resume, and an opt-in host memory-pressure backpressure guard. Tmuxlet is an open-source project distributed via GitHub, currently in active 0.5.x development, designed specifically as a dependency-free orchestration harness for terminal AI workloads.

Tags: cli, developer tools, OpenSource, productivity, terminal

Visit Tmuxlet

Key features

• Multiplexed tmux isolation — Launches and isolates supported CLI coding agents in uniquely targeted, exact-matched tmux sessions. - Machine-wide concurrency limits — Caps concurrent agent runs via a crash-safe live semaphore configurable with --max-concurrency or TMUXLET_MAX_CONCURRENCY. - Automated session reaper — Sweeps hung, leaked, or blocked agent sessions using tmuxlet reap with --blocked and --older-than flags. - Paced input broadcasting — Dispatches keystrokes and commands across all live sessions steadily with tmuxlet send-all to prevent system spikes. - Memory pressure protection — Throttles session spawning with an opt-in --memory-pressure-guard under heavy host RAM load. - Flexible prompt ingestion — Accepts prompts via CLI arguments, interactive flags (--prompt-interactive), or piped standard input via bare -.

Use cases

Use Case: Parallel Coding Agent Automation Run dozens of concurrent refactoring or auditing subtasks across multiple repositories without risking system lockups from unthrottled process creation.

Use Case: Automated Session Cleanup and Maintenance Incorporate tmuxlet reap into cron jobs or CI pipelines to automatically purge leaked agent processes and reclaim memory from abandoned coding sessions.

Use Case: Coordinated Multi-Agent Control Broadcast synchronized instructions or confirmations across multiple active agent instances in parallel using paced dispatch with tmuxlet send-all.

Who it is for

Tmuxlet is designed for developers, platform engineers, and AI workflow builders running multiplexed, terminal-based AI coding agents at scale. It is not suitable for developers seeking a single-instance GUI coding assistant or those who do not use tmux.

Install guide

Installation Clone the repository and install Tmuxlet into your local path (requires tmux to be installed): bash git clone https://github.com/CodefiLabs/tmuxlet.git cd tmuxlet # Symlink or copy the executable to your PATH sudo ln -s "$(pwd)/src/tmuxlet" /usr/local/bin/tmuxlet

First Run Verify that tmux is available and check Tmuxlet concurrency settings: bash tmux -V tmuxlet --help Optionally configure your concurrency limit via environment variable: bash export TMUXLET_MAX_CONCURRENCY=12

Example Session Dispatch a prompt to an agent via standard input, inspect active runs, and reap stale sessions: bash # Send a prompt to an agent target via stdin echo "Audit dependencies in package.json and create a report" | tmuxlet run -- claude - # Broadcast a confirmation keystroke to all running sessions tmuxlet send-all "y\n" # Sweep sessions older than 2 hours or blocked on user input tmuxlet reap --older-than 2h --blocked

Part of CLI Coding Agents

Categories: AI Coding Agents, Developer Tools, Productivity

Related CLI agents

• Cate — Cate is an open-source agent harness and orchestrator developed by 0-AI-UG. Built on TypeScript, Vite, and Electron, it provides an…
• Iris TUI — Iris TUI is an open-source terminal dashboard and harness for monitoring and managing coding agent sessions, developed by itzenata and…
• Playwright Praman — Playwright Praman is an open-source test harness and agent orchestrator developed by Maheshwar Kanitkar. It is designed to bridge terminal…
• wt — wt is an open-source CLI harness and workspace orchestration tool maintained by Sahil Ahuja on GitHub. It is built to…
• agent-trace — agent-trace is an open-source command-line transcript utility created by ertygiq. It is designed to inspect, parse, and export execution logs…
• Shell GPT — Shell GPT (sgpt) is an open-source command-line productivity tool created by Farkhod Sadykov (TheR1D). It connects standard terminal workflows directly…