🚀
Pop open a coding env
60 seconds. Pick a template for a Go service, a Next app, or an existing repo. The environment is ready before you finish your coffee.
Spin up an environment. R1 is the default agent; Claude Code and Codex are pre-installed too. Close the window, close your laptop, close your brain. They keep working. Reopen from anywhere. State intact.
CodeR1 is for engineers who want an agent workspace that survives the rest of their day: start an environment, hand it work, walk away, then reopen it later without rebuilding context.
60 seconds. Pick a template for a Go service, a Next app, or an existing repo. The environment is ready before you finish your coffee.
Plain English works. “Add rate-limit middleware to /auth.” R1 plans, executes, verifies, and ships while Claude Code and Codex stay available inside the same box.
Lid down. Train ride. Dinner. The environment keeps running. State lives with the environment, not with the browser window you happened to close.
Phone, tablet, hotel computer. Rejoin the same environment with the same files, console history, and pending work. No “where was I?” tax.
The product is the same. The reason to care changes depending on whether you are shipping alone, juggling multiple agents, or working under a stricter review bar.
You are the whole engineering team. You need an environment that can keep working on your behalf while you do everything else a company asks of one person. CodeR1 gives you persistent environments, mobile check-ins, and no infra to babysit.
You already know how to prompt an agent. The problem is coordination. CodeR1 gives you a grid of every environment, persistent state across all of them, and the option to swap between R1, Claude Code, and Codex inside the same workspace.
Some teams need coding help but cannot accept invisible model traffic or unaudited code generation. CodeR1 ships with RelayGate inline, R1 as the default agent harness, and receipts on every change so review starts from evidence instead of trust.
On desktop, CodeR1 shows you your running environments as a grid. You don't tab-switch to check on an agent; you glance across the screen. One is running tests. One is pairing on a migration. One has a PR ready for your review. Your attention moves, not your cursor.
The point is not another prompt window. The point is an environment that starts quickly, keeps context between sessions, and lets you switch agent families without rebuilding the box.
Pick a template, boot the environment, and land in a workspace that already knows how to run code, tests, and model calls.
State outlives every window. Disk keeps syncing. The environment keeps running whether you are on a laptop, in a meeting, or on a train.
R1 is the default harness. Claude Code and Codex are ready too. RelayGate governs every model request and receipts land on every change.
Mobile is on the roadmap as a responsive web surface, not a native app. The desktop dashboard is the shipping product today; the mobile-tuned layout below is a design preview. When it lands, environments render as toggleable side panels, terminal interactions are tuned for touch, and file browsing respects your thumb.
Rendered preview. Full mobile experience is the real product; try it from your phone after you sign up.
Tap an environment to enter its console or files. The same environments from your desktop grid, sized for a thumb.
The appeal is not cinematic. It is the ordinary rhythm of letting long-running work continue while you move between review, meetings, and the rest of your day.
CodeR1 is for engineers who want the agent and the environment to survive every reconnect. That is a different category from local IDE copilots and a different tradeoff from generic hosted dev boxes.
Want the deeper mechanics? See how managed vs. BYO model billing works or walk a real request through RelayGate inline.
Environments run on CodeR1 infrastructure, not on your device. When you close a browser tab, close your laptop, or lose your connection, the environment keeps running. When you reopen from any device, state is intact. Scroll through a real disconnect-reconnect cycle.
The environment ran for 6h 40m across 3 devices and 2 disconnects. State loss: zero.
You can start solo, bring a team later, or jump straight to a governed rollout. The full matrix lives on /pricing.
$20/mo. Bring your own keys or use the managed pool.
Shared workspaces, shared review loops, same environment model.
SSO, audit, policy, and sovereign deployment options.
Managed inference is metered by token when you use it. Bring-your-own keys keep your provider bill on your side. Full detail lives on /pricing.
Start a trial if you want to feel the persistence model yourself. Book a walkthrough if you want to see the grid, the review loop, and the governed path with your own workflows in mind.
Open one environment, hand it work, close the laptop, and reopen later from somewhere else.
30 minutes. We show the grid, the persistence model, the review loop, and how RelayGate fits into the same environment.