Whats 1Code? An app to run your Claude Code agents in parallel that works on Mac and Web. On Mac - run locally, with or without worktrees. On Web - run in remote sandboxes with live previews of your app, mobile included, so you can check on agents from anywhere. Running multiple Claude Codes in parallel dramatically sped up how we build features.
Hi It’s Serafim and Sergey from 21st.dev, in the last 10 months we’ve built and launched 9 products, getting over 1 millions total users, $200k in revenue, and 10k GitHub stars.
Motivation
Claude Code has been our go-to for 4 months. When Opus 4.5 dropped, parallel agents stopped needing babysitting.
You could finally run multiple and trust them. But the more agents we ran, the more the CLI felt like a limitation, not a feature.
So we built 1Code.dev
What’s next:
Bug bot for identifying issues based on your changes; QA Agent, that check that new features dont break anything
Adding OpenCode, Codex, other models and coding agents.
API for starting Claude Codes in remote sandboxes. Debugging and spotting errors in AI Coding Era is more critical than ever
Try it:
We're open-source, so you can just bun build it.
Paid tiers exist too - for us it’s gonna be highest signal that we're on the right track.
Either way, we'd love to hear your feedback!
How do you usually decide what tasks to split across agents versus keeping in a single workflow?
This is wild. I've tried probably 10 different AI coding tools and they all fall apart when you're working across multiple files or trying to understand existing codebases. The fact that this actually reads your whole project context and suggests changes that don't break everything else is exactly what's been missing.
One thing I'm curious about—does it handle refactoring well? Like if I want to rename a function that's used in 20 files, will it catch all the references or do I still need to double-check manually?
Also the $30/month unlimited is actually reasonable compared to Cursor/Copilot pricing. Gonna try this out.
Parallel agents are the real speed unlock. GTWY is built around the same idea: let AI operate concurrently, but predictably, inside real systems.
Great app, love Claude Code, but this is just better for non technical people and even for me when I’m too distracted to open IDE.
Can you add preview?
Congrats on the launch
Is this Claude Code Agent or its another agent that uses Agent SDK, Opus 4.5 with Claude Subscription?
Wow, 1Code looks incredible! The live preview feature for mobile apps while running agents is a game changer. How does the remote sandboxing handle file access for more complex projects?
The parallel agent problem so real. Curious how you're handling conflict resolution when agents touch the same files?
If a user already has Cursor (or another agentic editor) plus a couple terminal tabs, what’s the strongest reason they’d switch to 1Code instead of staying in their editor—what workflow becomes materially faster or safer only when you add a dedicated multi-agent “control plane”?
9 products, 1 million users, $200k revenue... And I still can't figure out how to fix my Wi-Fi. :D But seriously, 1Code looks super amazing!
That's crazy. I've just given in and bought Claude code today after an infuriating few days with Cursor. Feels like Cursor is dead unfortunately, it was a great run. Very excited to try this out, hope you're going to maintain and push this forward!
Finally something to not be scared of worktrees
Congrats on the launch!
Sounds really cool! Gonna check it out and return with a feedback
GL with your launch 🤟
Running multiple Claude Code agents in parallel is a massive unlock for velocity compared to serial chats. How does the system handle conflict resolution if two agents try to edit the same file structure simultaneously?
