ntm

ntm (Named Terminal Multiplexer / Agent Harness) is an open-source terminal multiplexer orchestrator created by Jeffrey Emanuel (Dicklesworthstone) and designed specifically for orchestrating multi-agent AI coding workflows. Built primarily in Go, ntm manages multi-pane terminal environments to coordinate and supervise autonomous coding agents executing tasks in parallel. In day-to-day use, ntm operates as a terminal-first harness that provisions dedicated multiplexer panes for various LLM agents—including Claude, Codex, GPT variants, and Gemini. It orchestrates agent lifecycles, assigns durable pane identities before launch, and enables structured multi-model exploration across parallel sessions. Developers interact via CLI commands and configurable command palettes defined in markdown to trigger prompt refinements and monitor agent outputs. What distinguishes ntm from single-agent CLI harnesses is its focus on multi-agent terminal coordination, integrated Active Prompt Refinement (.apr) workflows, and robust pane identity tracking across sessions. Rather than running a single agent sequentially, ntm provides the scaffolding to run ensemble tasks, hypersync specs, and multi-model parallel designs across isolated pane sessions. ntm is free and open-source software hosted on GitHub, with active community contributions and continuous testing pipelines for integration, end-to-end multi-agent suites, and plugin presets.

Tags: agent harness, cli, coding agent, developer tools, terminal

Visit ntm

Key features

  • Multiplexer pane orchestration — Manages and automates dedicated terminal panes for concurrently running AI coding agents. - Durable pane identities — Publishes and coordinates deterministic identities for agent panes prior to process launch to ensure robust lifecycle tracking. - Active Prompt Refinement (.apr) — Native support for multi-round prompt workflows and oracle inputs to iteratively refine task instructions. - Multi-model parallel design — Supports side-by-side execution and spec evaluation across multiple models including Claude Opus, GPT, and Gemini. - Command palette integration — Reads runtime command palettes from localized markdown files to streamline frequent multi-pane developer actions. - Plugin and preset ecosystem — Includes maintainer-curated presets (such as Oh My Pi) and integration hooks for extended workflow automation.

Use cases

Use Case: Multi-Model Parallel Architecture Evaluation A software architect needs to compare implementation strategies for a new service across multiple frontier LLMs. Using ntm, they spawn isolated terminal panes running Claude, Gemini, and GPT instances simultaneously against the same specification to evaluate differing architectural approaches side by side.

Use Case: Active Prompt Refinement for Agent Tasks A developer wants to ensure complex refactoring tasks do not fail due to ambiguous requirements. They use ntm's .apr workflow harness to iterate on prompt specifications across structured review rounds before distributing final execution commands across agent panes.

Use Case: Multi-Agent Workspace Orchestration A team lead runs an ensemble of specialized coding agents handling linting, test generation, and code implementation in parallel. ntm coordinates the terminal panes and tracks durable agent identities to prevent race conditions and maintain observable logs.

Who it is for

ntm is intended for power developers, systems programmers, and AI engineers who orchestrate multiple CLI coding agents concurrently across terminal multiplexers. It is ideal for users comfortable with tmux workflows, Go tooling, and multi-model terminal pipelines. It is less suitable for developers seeking a single-pane GUI chat interface or a zero-configuration IDE extension.

Install guide

Installation Clone the repository and build the Go binary: bash git clone https://github.com/Dicklesworthstone/ntm.git cd ntm go build -o ntm ./... sudo mv ntm /usr/local/bin/

First Run Verify the installation and review available orchestrator options: bash ntm --help Ensure required AI CLI agent binaries (such as Claude CLI, Codex tools, or custom runner scripts) and your terminal multiplexer dependencies are installed in your PATH.

Example Session Launch an orchestrated session with a configured command palette or workflow: bash # Initialize an ntm session in your project root ntm session start --name dev-ensemble # Spawn managed agent panes according to your workspace configuration ntm spawn --config .apr/workflows/hypersync.yaml

Part of CLI Coding Agents

Categories: AI Coding Agents, Code Assistants, Developer Tools

Related CLI agents

  • Patapim — Patapim is an agent harness and orchestration environment built to wrap and extend CLI coding agents like Anthropic's Claude Code,…
  • Repomon — Repomon is an open-source developer harness and repository companion tool created by Ali Hamza Azam. Built with a Rust core…
  • Unpeel — Unpeel is a terminal multiplexer and agent orchestration harness designed for managing persistent developer tasks and AI coding agents. It…
  • Even — Even is an agent-native desktop workspace and orchestration environment developed by Todience, Inc. It integrates a native terminal, an embedded…
  • AgentCliMenu — Agent CLI Menu (binary `agent-cli-menu`, alias `acm`) is an open-source session manager and agent harness developed by Roy Padina. Designed…
  • Showagent — Showagent is an open-source terminal user interface (TUI) harness and session manager created by aytzey for developers who run multiple…
  • AI Tools
  • Categories
  • Industries
  • CLI Coding Agents
  • MCP Servers
  • MCP Categories