generalaction/emdash

Emdash – Parallel AI Coding Agent Desktop App

What it is: Emdash is a native desktop application (macOS, Windows, Linux) that lets you run multiple AI‑powered coding assistants (Claude Code, Codex, GitHub Copilot, etc.) at the same time, each isolated in its own Git worktree. It provides a single UI for sending tickets, reviewing diffs, creating PRs and merging results, and it works on local repositories as well as codebases on remote machines accessed via SSH/SFTP.

Key capabilities

• Parallel agents – launch any number of supported AI coding agents without opening separate terminals.
• Git worktree isolation – every agent works in its own worktree/branch, so you can explore alternative fixes or features side‑by‑side and compare the generated diffs.
• Ticket integration – import issues from Linear, GitHub, Jira, GitLab, Asana, Monday.com, and other trackers directly into an agent session.
• Remote execution – connect to a remote host over SSH/SFTP; the same parallel workflow runs on the remote codebase, with SSH keys or certificates stored in the OS keychain.
• Provider‑agnostic – detects installed CLI tools for Claude Code, Codex, Cursor, OpenCode, Amp, Devin, Qwen Code, Droid, GitHub Copilot, etc. Agents that support lifecycle hooks get automatic status tracking.
• Local‑first privacy – all state lives in a local SQLite DB; Emdash itself never uploads code or chat logs. Only the underlying agent CLIs may send data to their providers.
• Optional telemetry – can be turned off via a setting or the TELEMETRY_ENABLED=false environment variable.

Installation (pre‑built binaries are provided)

Platform	Install command / download
macOS	brew install --cask emdash (or download the ARM64/Intel DMG)
Windows	MSI installer or portable .exe from the releases page
Linux (x64)	AppImage, DEB, or RPM from the releases page
Linux (ARM64)	AppImage, DEB, or RPM for ARM64

Typical workflow

1. Open Emdash and add a project (local folder or remote SSH host).
2. Choose a coding agent (e.g., Claude Code) and create a new task, optionally pulling an issue from your tracker.
3. The app creates a fresh Git worktree/branch, launches the agent CLI, and streams its suggestions.
4. Review the generated diff inside the UI, run CI checks, and merge the change if it looks good.
5. Repeat with other agents in parallel to compare different solutions.

Who it’s for

• Developers who use AI coding assistants and want a cleaner, multi‑agent experience.
• Teams that need to triage many tickets quickly, turning each into an isolated AI‑generated fix.
• Users who prefer a local‑first tool that does not expose their source code to a third‑party server.

License: Apache‑2.0 (open source).

Links: Download page • Documentation • Discord community • Y Combinator W26 demo

All details are taken directly from the repository’s README.