l0ng-ai/tty7

The terminal that outlives its window. Shells survive quitting the app, agents can drive other agents, and remote machines work the same way. Pure Rust, GPU-rendered.

What it solves

tty7 is a terminal workbench designed to solve the fragility and lack of integration between terminal emulators and AI coding agents. It eliminates the need for external session managers like tmux by providing persistent sessions that survive app quits and reboots, while creating a first-class environment for AI agents to operate within the CLI.

How it works

A background server manages shells and panes independently of the GUI window. It is built using Rust, utilizing Zed's gpui for GPU rendering and Alacritty's VT core. This architecture allows for remote development via a native SSH stack and ensures that sessions, layouts, and agent conversations remain active even when the client is disconnected.

Who it’s for

Developers and power users who heavily utilize AI agents (such as Claude Code, Codex, or Copilot) in their terminal and require a high-performance, persistent workspace that supports remote development.

Highlights

  • Agent-Awareness: Automatically detects 28 different AI agents, providing status indicators, notifications, and git context for every repository.
  • Agent Scriptability: Includes a bundled CLI and "skills" that allow one AI agent to open panes, hand off tasks to other agents, and capture results.
  • High Performance: Outperforms popular terminals like Alacritty and Ghostty in plaintext I/O and frame rates.
  • Editor-Grade Input: Features ghost suggestions from history, syntax highlighting, and multi-line editing without requiring plugins.
  • Remote Workspaces: Native SSH stack supporting profiles, keychain secrets, SFTP, and port forwarding, keeping data on the remote machine.

Related

  • Project
    remorses/tuistory

    tuistory is a Node‑JS CLI and library that runs any terminal command in a background PTY session, letting AI agents read logs, wait for output, type, press keys, capture screenshots, and restart the process. Humans can attach to the same session to see exactly what the agent sees. It replaces ad‑hoc `tmux`+`sleep` scripts with deterministic, event‑driven control for AI‑driven automation.

  • Project
    nowledge-co/con-terminal

    con is a Rust‑based, GPU‑accelerated terminal emulator that includes a built‑in AI agent for context‑aware command assistance. It works on macOS (beta), Windows (early beta) and Linux (preview), supports typical CLI workflows (ssh, tmux), and offers configurable shortcuts, a `con-cli` helper, and extensible “skills”. Install via Homebrew, curl scripts, or platform‑specific package managers. MIT‑licensed.

  • Dispatch
    hrdx Terminal Multiplexer

    hrdx is a lightweight terminal multiplexer written in Go, specifically designed to manage multiple AI coding agents and shell sessions across various workspaces.

  • Project
    iAmCorey/kooky

    A minimal, GPU-accelerated terminal for macOS designed for AI coding, featuring one-click agent launchers, workspace management, and integrated SSH/Git worktree support.

  • Project
    rengwu/chartr

    A terminal-based agentic workspace that organizes CLI agents, terminals, and tools into persistent, project-specific spaces.