Tmux
MCP ServerFree** - Interact with your Tmux sessions, windows and pane, execute commands in tmux panes and retrieve result.
Capabilities10 decomposed
mcp-based tmux session enumeration and introspection
Medium confidenceExposes tmux session hierarchy through MCP resource protocol, allowing AI assistants to discover and inspect all active sessions with metadata including session names, window counts, creation timestamps, and attachment status. Implements resource subscription pattern via @modelcontextprotocol/sdk to enable real-time session state synchronization and dynamic resource updates when sessions are created or destroyed.
Implements MCP resource protocol for tmux introspection rather than simple command wrapping, enabling Claude Desktop to maintain a persistent view of session state through resource subscriptions and change notifications. Uses tmux list-sessions with format strings to extract structured metadata without parsing text output.
Provides standardized MCP integration for Claude Desktop whereas shell scripts or REST APIs require custom integration work; resource-based architecture enables real-time state awareness vs polling-based alternatives.
asynchronous command execution with deferred result retrieval
Medium confidenceExecutes shell commands in tmux panes without blocking the MCP server, returning a command ID immediately and allowing result retrieval via separate resource lookup. Implements fire-and-forget execution pattern with optional polling via tmux://command/{commandId}/result resources, supporting both synchronous workflows (wait for completion) and asynchronous patterns (fire and check later). Handles shell-specific exit code detection through configurable --shell-type parameter to correctly identify command success/failure across bash, zsh, and fish.
Decouples command execution from result retrieval through MCP resource protocol, enabling non-blocking execution patterns where the AI assistant can fire commands and poll results independently. Uses shell-specific exit code markers (e.g., echo $? for bash) to reliably detect command completion and success status across different shell environments.
Provides true asynchronous execution with deferred result retrieval vs synchronous SSH/exec alternatives that block until completion; shell-type configuration ensures accurate exit code detection across heterogeneous environments vs generic command wrappers that assume single shell type.
interactive pane content capture with color preservation
Medium confidenceCaptures the current visible content of a tmux pane with optional ANSI color code preservation, enabling AI assistants to read terminal output including colored text, syntax highlighting, and styled formatting. Implements configurable capture modes via capture-pane tool that can preserve raw ANSI escape sequences or strip them for plain text, supporting both human-readable colored output and machine-parseable plain text depending on use case. Handles pane history buffer retrieval to capture scrollback content beyond the visible viewport.
Provides dual-mode capture (colored vs plain text) via single tool interface, allowing AI assistants to choose between human-readable colored output and machine-parseable plain text. Uses tmux capture-pane with -p (print) and -S (start line) flags to efficiently retrieve both visible viewport and scrollback history without spawning separate processes.
Preserves ANSI color codes for semantic understanding vs plain text alternatives that lose formatting context; supports scrollback history retrieval vs simple screen capture that only shows visible content.
pane lifecycle management with split and kill operations
Medium confidenceCreates, splits, and destroys tmux panes within windows through MCP tools, enabling AI assistants to dynamically manage terminal layout and organize command execution across multiple panes. Implements split-pane operation with configurable split direction (horizontal/vertical) and target pane selection, allowing creation of new panes for parallel execution. Supports pane destruction via kill-pane tool with optional confirmation to prevent accidental data loss.
Exposes tmux pane splitting and killing as MCP tools with structured input/output, enabling AI assistants to programmatically manage terminal layout without shell command knowledge. Uses tmux split-window and kill-pane commands with format string parsing to return new pane identifiers for subsequent operations.
Provides structured pane management vs manual tmux commands that require shell knowledge; enables dynamic layout creation during AI workflows vs static pre-configured layouts.
raw mode command execution for interactive applications
Medium confidenceExecutes commands in tmux panes with raw mode enabled, allowing interactive applications like REPLs, text editors, and TUI tools to receive input and maintain state across multiple interactions. Implements key injection without automatic Enter appending, enabling navigation of interactive menus and TUI applications through arrow keys and special characters. Maintains pane state between command invocations, allowing AI assistants to interact with long-running interactive sessions (Python REPL, Node REPL, vim, etc.).
Supports raw mode execution with key injection without Enter, enabling stateful interaction with interactive applications vs simple command execution that assumes line-based input. Maintains pane state across multiple invocations, allowing AI assistants to build multi-turn conversations with REPLs and interactive tools.
Enables interactive REPL workflows vs batch command execution that cannot maintain state; key injection without Enter supports TUI navigation vs line-based alternatives limited to simple commands.
window lifecycle management with creation and destruction
Medium confidenceCreates and destroys tmux windows within sessions through MCP tools, enabling AI assistants to organize command execution across multiple windows within a single session. Implements window creation with optional command execution in the new window, allowing immediate setup of new windows for specific tasks. Supports window destruction via kill-window tool with proper cleanup of all contained panes.
Exposes tmux window creation and destruction as MCP tools with structured input/output, enabling AI assistants to organize workflows across multiple windows without shell command knowledge. Uses tmux new-window and kill-window commands with format string parsing to return window identifiers.
Provides structured window management vs manual tmux commands; enables dynamic window creation during workflows vs static pre-configured layouts.
session lifecycle management with creation, discovery, and termination
Medium confidenceCreates, discovers, and destroys tmux sessions through MCP tools and resources, enabling AI assistants to manage the top-level session hierarchy. Implements session creation with optional initial command and window setup, session discovery via list-sessions and find-session tools with metadata extraction, and session termination via kill-session. Uses tmux list-sessions with format strings to extract structured metadata (session name, window count, creation time, attachment status) without text parsing.
Implements MCP resource protocol for session discovery with structured metadata extraction via format strings, enabling AI assistants to maintain awareness of session state without text parsing. Supports session creation with initial command setup, allowing immediate task execution in new sessions.
Provides structured session management vs manual tmux commands; format string-based metadata extraction is more reliable than text parsing for session discovery.
mcp protocol server implementation with stdio transport
Medium confidenceImplements a complete MCP server using @modelcontextprotocol/sdk that exposes tmux functionality through standardized MCP primitives (tools, resources, prompts). Operates as a Node.js process communicating with Claude Desktop via stdio transport, translating MCP protocol requests into tmux commands and returning structured responses. Declares server capabilities including resource subscription support, tool change notifications, and logging, enabling dynamic resource updates and real-time state synchronization.
Implements full MCP server specification with resource subscription support and capability declaration, enabling Claude Desktop to maintain persistent awareness of tmux state. Uses stdio transport for communication, allowing seamless integration with Claude Desktop's MCP client without network configuration.
Provides standardized MCP integration vs custom Claude plugins that require separate maintenance; resource subscription enables real-time state awareness vs polling-based alternatives.
shell-specific command execution with exit code detection
Medium confidenceExecutes commands with shell-specific configuration to reliably detect command completion and exit status across bash, zsh, and fish environments. Implements configurable --shell-type parameter that adjusts exit code detection mechanisms based on shell environment, using shell-specific markers (e.g., echo $? for bash) to identify command success/failure. Handles shell-specific quirks like different prompt formats and exit code variable names, ensuring accurate command result detection regardless of shell configuration.
Implements shell-specific exit code detection via configurable --shell-type parameter, enabling reliable command result detection across heterogeneous shell environments. Uses shell-specific markers and prompt detection rather than generic timeout-based completion detection.
Provides accurate exit code detection across multiple shells vs generic command wrappers that assume single shell type; shell-specific configuration ensures reliability vs timeout-based completion detection that may fail with slow commands.
claude desktop configuration integration via npm package
Medium confidenceDistributes tmux-mcp as an npm package that Claude Desktop can install and configure through its mcpServers configuration file, enabling one-command setup without manual compilation or path configuration. Implements binary distribution model where TypeScript source is compiled to JavaScript and exposed as tmux-mcp command, allowing Claude Desktop to invoke the MCP server directly. Supports configuration parameters (--shell-type) passed through Claude Desktop's mcpServers config, enabling environment-specific customization without code changes.
Distributes as npm package with pre-compiled JavaScript binaries, enabling Claude Desktop to invoke tmux-mcp directly without TypeScript compilation. Supports configuration parameters through Claude Desktop's mcpServers config, allowing environment-specific customization without code changes.
One-command npm installation vs manual compilation and path configuration; pre-compiled binaries enable immediate use vs source distribution requiring build step.
Capabilities are decomposed by AI analysis. Each maps to specific user intents and improves with match feedback.
Related Artifactssharing capabilities
Artifacts that share capabilities with Tmux, ranked by overlap. Discovered automatically through the match graph.
agent-of-empires
Manage multiple Claude Code, OpenCode agents from either TUI or Web for easy access on mobile. Also supports Mistral Vibe, Codex CLI, Gemini CLI, Pi.dev, Copilot CLI, Factory Droid Coding. Uses tmux and git worktrees.
mcphub.nvim
** A Neovim plugin that provides a UI and api to interact with MCP servers.
MCPHub
** – An Open Source macOS & Windows GUI Desktop app for discovering, installing and managing MCP servers by **[Jeamee](https://github.com/jeamee)**
mcp-client-for-ollama
A text-based user interface (TUI) client for interacting with MCP servers using Ollama. Features include agent mode, multi-server, model switching, streaming responses, tool management, human-in-the-loop, thinking mode, model params config, MCP prompts, custom system prompt and saved preferences. Bu
MBro
** - A powerful interactive terminal **M**CP **Bro**wser client with tab completion and automatic documentation that allows you to work with multiple MCP servers, manage tools, and create complex workflows using AI assistants.
gptme
Your agent in your terminal, equipped with local tools: writes code, uses the terminal, browses the web. Make your own persistent autonomous agent on top!
Best For
- ✓AI-assisted terminal workflow automation
- ✓Teams using Claude Desktop for DevOps tasks
- ✓Developers building AI agents that manage multiple terminal sessions
- ✓Long-running build/test workflows in CI/CD contexts
- ✓Multi-pane parallel execution scenarios
- ✓AI agents that need non-blocking command execution to maintain responsiveness
- ✓Reading colored terminal output (test results, logs, syntax-highlighted code)
- ✓Capturing REPL/editor state for AI analysis
Known Limitations
- ⚠Requires tmux to be installed and running on the host system
- ⚠Session discovery is limited to the user's tmux namespace — cannot access other users' sessions
- ⚠Resource subscription updates depend on MCP client polling frequency, not true event-driven updates
- ⚠Result retrieval requires polling via resource lookup — no webhook or event-driven callbacks
- ⚠Command timeout handling is implicit (relies on shell timeout mechanisms), not explicitly managed by MCP server
- ⚠Exit code detection accuracy depends on shell configuration and --shell-type parameter correctness
Requirements
Input / Output
UnfragileRank
UnfragileRank is computed from adoption signals, documentation quality, ecosystem connectivity, match graph feedback, and freshness. No artifact can pay for a higher rank.
About
** - Interact with your Tmux sessions, windows and pane, execute commands in tmux panes and retrieve result.
Categories
Alternatives to Tmux
Are you the builder of Tmux?
Claim this artifact to get a verified badge, access match analytics, see which intents users search for, and manage your listing.
Get the weekly brief
New tools, rising stars, and what's actually worth your time. No spam.
Data Sources
Looking for something else?
Search →