Hey YC! 👋 We’re Arne and Raban the we’re building Emdash. An open-source desktop app that lets you work with multiple coding agents at once. We’re calling it an Agentic Development Environment (ADE).

https://youtu.be/jdIPf2cxf6w

💬 Problem
Every time we used a coding agent, we'd start a task and ... wait. We tried running multiple agents on the same repo and ended up with branch chaos and merge conflicts. When a new model dropped – Opus, Codex, Gemini flash, Kimi, whatever – we had no clean way to test it against our current setup on the same task. We were bouncing between terminals, GitHub, and our IDE.

⏩ Solution

Emdash is an environment for agentic development that lets you run multiple agents side-by-side safely, review what each one did in one place, and create a PR when it looks right.

🔍 How it works

• Start a task → Emdash spins up a new git worktree so nothing touches your main directory
• Choose any agent, out of 20+ providers(incl. Claude Code, Codex, Gemini)
• Run multiple agents at once without conflicts
• Compare diffs, make quick manual edits if needed, and create a PR

🧰 What’s included

• Worktree isolation per agent
• Issue handoff from Linear / Jira / GitHub
• One place to review diffs and ship PRs
• Lifecycle scripts + CI/CD hooks
• Best-of-N comparisons
• Integrated file editor + browser preview
• Remote project access via SSH

We natively embed ~20 agent providers (including Claude Code, Codex, Continue, and Codebuff), so you can use the latest models inside the same workflow.

To get started run brew install –cask emdashor download directly from GitHub releases. We’re available on macOS and Linux.

🙏 Our Asks

1. Try Emdash - it’s free and open-source
2. Give us a star on GitHub 🌟

Questions or feedback? Join our Discord or email us at founders@emdash.sh

— Arne & Raban