robzilla1738

The native macOS terminal that keeps your sessions running and tells you when a coding agent needs you. GPU-rendered, scriptable, agent-aware.

29
0
89% credibility
Found Jun 01, 2026 at 29 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Swift
AI Summary

Harness is a native macOS terminal application that renders text using the computer's graphics chip, keeps your terminal sessions alive even after closing the app, and automatically detects AI coding assistants to notify you when they need your attention or approval.

How It Works

1
💻 Download and install Harness

You download the app from GitHub, drag it to your Applications folder, and launch it like any other Mac app.

2
🚀 See your first terminal window

The app opens with a clean, fast terminal that looks great right out of the box with a muted dark theme.

3
🔄 Your work survives closing the app

You close Harness, reopen it later, and all your tabs and running commands are exactly where you left them.

4
🤖 Run an AI coding assistant

You start Claude Code, Cursor, or another AI agent to help with your project — Harness automatically notices it.

5
The agent needs your attention
📱
Desktop notification

A banner appears on your screen with the agent's name and what it needs

🔴
Visual ring in the tab

The tab with the waiting agent gets a colored ring so you can spot it instantly

6
⌨️ Use keyboard shortcuts to navigate

You press Cmd+Shift+U to jump straight to the waiting agent, or use Cmd+K to open the command palette for quick actions.

Seamless AI-assisted development

Your terminal keeps everything running, watches your agents, and alerts you at the right moment — so you can focus on building.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 29 to 29 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is harness-cli?

Harness is a native macOS terminal built in Swift that keeps your sessions alive even after you quit the app. It runs a background daemon so your tabs and splits survive restarts, and you can attach to the same session from a second window or over SSH. The terminal renders via a custom GPU engine, supports 490 color themes, and—most notably—it watches the AI coding agents running inside it (Claude Code, Codex, Cursor, and others) and alerts you the moment one needs your attention.

Why is it gaining traction?

The killer feature is agent awareness. Most terminal emulators treat AI tools like any other process. Harness actively monitors them, shows which tab is running what, and pings you with a desktop notification when an agent stops or needs approval. A single `Cmd+Shift+U` jumps straight to whoever is waiting—skipping the tabs still generating output. For developers running agents across multiple projects, this is a genuine quality-of-life improvement over constantly alt-tabbing to check progress.

The CLI (`harness-cli`) is equally capable: send-keys, capture-pane, resize-pane, swap-pane, and full session management all work from the command line. The daemon architecture means your layout persists without a config file in sight.

Who should use this?

Developers using AI coding assistants who run multiple agents across different projects will get the most value. macOS power users who want session persistence without tmux configuration overhead. Anyone who wants a GPU-rendered, scriptable terminal that works out of the box on Apple silicon.

Verdict

At 29 stars this is early-stage, but the engineering is solid—Swift 6, comprehensive tests, CI/CD, signed releases, and thorough documentation including a tmux-style capability PDF. The credibility score sits around 0.9% which reflects the project's newness, not its quality. If you live in a terminal and use coding agents, this is worth a weekend test drive. Just know you're adopting before the community has fully stress-tested it in production.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.