Emdash

Emdash is an open-source agentic development environment (ADE) from Y Combinator's Winter 2026 batch. It runs multiple coding agents in parallel — Claude Code, Codex, OpenCode, Amp, and more — each isolated in its own Git worktree with its own branch, terminal, and conversation. Pull issues from Linear or GitHub, review diffs, and merge PRs from one window, locally or over SSH.

Key Features

• Parallel agents in Git worktrees: Every task gets an isolated worktree and branch, so agents work on the same codebase without collisions.
• Provider-agnostic: Auto-detects installed CLIs — Claude Code, Codex, Cursor, OpenCode, Amp, Devin, Qwen Code, and more (25+ supported).
• Ticket-to-PR pipeline: Assign issues from Linear, GitHub, Jira, GitLab, Asana, and more; review diffs, open PRs, check CI, and merge in one place.
• SSH remote execution: Run the same parallel workflow on remote machines with SSH/SFTP, credentials stored in your OS keychain.
• Fast task startup: Background worktree reserve brings task creation down to ~500-1000ms.
• Full dev loop: Built-in terminals, diff viewer, browser preview, prompts, skills, and MCP tool management.

Use Cases

Who Should Use This Tool?

• Multi-agent orchestrators: Developers who want a whole team of agents working in parallel without terminal chaos.
• Ticket-driven teams: Engineers who feed Linear/Jira/GitHub issues straight to agents.
• Remote dev setups: Teams running agents where the code lives, over SSH.

Problems It Solves

1. Agents fighting over files: Git worktree isolation means two agents can edit the same repo without clobbering each other.
2. One-at-a-time bottlenecks: Parallel execution replaces the slow "run one agent, wait, repeat" loop.
3. Orchestration sprawl: Issues, diffs, PRs, and CI checks live in one window instead of five tools.

Pricing

Emdash is free and open source. Install via Homebrew (brew install --cask emdash) or download from the website. Your code and prompts stay local — nothing is sent to Emdash servers.

Advantages & Unique Selling Points

Compared to Competitors:

1. Worktree-native isolation: Uses git's native worktrees rather than a custom sandbox, so agents share one history without conflict.
2. Truly provider-agnostic: Unlike walled-garden tools, it works with whatever agent CLIs you already have.
3. Full loop, not just a launcher: Ticket intake, review, PR, and merge are included, not bolted on.

What Makes It Stand Out:

• YC W26 backing with rapid adoption: 800k+ downloads and 5k+ GitHub stars.
• Runs locally or on your own remote machines; data never leaves your infrastructure.
• Sub-second task startup via background worktree reserve.

User Reviews

"Point two coding agents at the same project folder, and they fight... Emdash gives every agent its own worktree. Agent one works here, agent two works there — same codebase, separate folders, zero collisions." — Use AI with Tech Dad, YouTube review

"Running multiple coding agents in parallel, each isolated in its own git worktree... Emdash puts the terminal at the center and makes it easy to run multiple agents at once." — Arne Strickmann, co-founder, Show HN

Getting Started

Quick Start Guide

1. Install: brew install --cask emdash or download from emdash.sh.
2. Open a project: Point Emdash at a local repo or connect over SSH.
3. Create a task: Assign an issue or type a prompt — Emdash spins up an isolated worktree.
4. Review and merge: Inspect the diff, open a PR, watch CI, and merge from the same window.

Integration

• Providers: Claude Code, Codex, Cursor, OpenCode, Amp, Devin, Qwen Code, Droid, GitHub Copilot, and more.
• Issue trackers: Linear, GitHub, Jira, GitLab, Asana, Monday.com, Forgejo, Plain, Featurebase.
• Remote: SSH/SFTP with agent, key, or password auth.

Frequently Asked Questions

Is Emdash open source?

Yes, fully open source under Apache 2.0, built in the open by Arne Strickmann and Raban von Spiegel (YC W26).

Does it lock me into one agent?

No. Emdash is provider-agnostic by design — it auto-detects installed agent CLIs and lets you mix providers per task.

Where does my code run?

On your machine or your own remote server over SSH. Emdash doesn't proxy your code or prompts to its servers.

Alternatives

If Emdash isn't the right fit, consider these alternatives:

• Herdr: Terminal-based runtime that keeps agents alive in persistent sessions.
• Superset: Run an army of CLI agents in isolated worktrees with automations.
• cmux: macOS terminal with vertical tabs and agent notifications for manual multitasking.

Tips & Best Practices

1. Mix providers by task: Use a cheaper or faster model for refactors and a stronger one for tricky logic.
2. Feed real tickets: Pass Linear/GitHub issues with acceptance criteria instead of loose prompts.
3. Use SSH for heavy work: Run agents on a powerful remote box and keep your laptop light.

Conclusion

Emdash turns you from author into orchestrator: parallel coding agents, each safely isolated in its own worktree, with the whole loop from ticket to merge in one window. If you're ready to run a team of agents instead of one at a time, it's the cleanest open-source entry point.