210k+ stars on GitHub
•
Built by Anomaly

Run OpenCode
‍
in the Cloud

One minute and your coding agent is live 24/7

Launch in Cloud

Atomic Bot

running OpenCode yourself

Atomic Bot
Included
Live in a minute, no install
Included
Runs 24/7, laptop off
Included
Our models or your own key
Included
One-click updates + backup
Included
Support from our team
OpenCode
OpenCode
Not included
Install script and a terminal
Not included
Only while your laptop is on
Not included
A key and a bill per provider
Not included
Updates and restarts on you
Not included
You fix it when it breaks

Why developers pick OpenCode

Everything that made OpenCode popular, now running in the cloud around the clock.

210k+stars on GitHub
Open source

MIT licensed and built in the open by 950 contributors. Over 16M developers use it every month.

Fix flaky testsWorking
Refactor API clientWorking
Write release notesDone
Parallel sessions

Start several sessions on the same project and let them work side by side. Each one keeps its own history.

BuildPlanTab
Build and Plan

Build writes and runs the code. Plan studies your project and proposes changes without touching a file. Both work right in the chat on Atomic Bot.

/undo/redo
Undo any step

Not happy with a change? Type /undo right in the chat on Atomic Bot to roll back the files and the conversation. In the TUI, /redo brings them back.

Hosting done for you

No servers, no terminal, no setup. Launch OpenCode in one click, and we keep it online, updated and ready.

Always on, 24/7

OpenCode lives in the cloud, not on your laptop. Close the lid, lose Wi-Fi or fly across the world, and it keeps working.

  • Keeps running when your computer is off
  • Long tasks keep going when you close the tab
  • Replies whenever you message it
OpenCode · activity overnight
00061218
24/7online
Your laptopLid closed
Asleep
Your agentAtomic Bot cloud
Working

One account, many agents

Run a whole team from one login: OpenCode on your codebase, OpenClaw on your inbox, Hermes on research. Or run 300 copies of OpenCode side by side.

  • Mix different agents in one account
  • Scale one agent to as many copies as you need
  • Each agent keeps its own model, memory and settings
Your agents
OpenCodeBackend refactor
OpenAIOnline
OpenClawInbox and calendar
AnthropicOnline
HermesMarket research
Z.aiOnline
Mercury AgentTeam knowledge base
DeepSeekOnline
NanoClawTelegram support bot
QwenOnline
One account, one login

100+ models from top labs

Pick the model for OpenCode from Anthropic, OpenAI, Google, DeepSeek, Qwen, Z.ai and more. One agent runs on one model, and you can switch it in a click.

  • No separate accounts or keys with each lab
  • Switch models without reinstalling anything
  • Open-weight and frontier models side by side
OpenCode · settings
OpenCodeOpenCode runs on Anthropic

One agent runs on one model. Pick another one and it switches right away.

Dollars, not tokens

Your dashboard shows spend in plain dollars and how long your balance lasts at your current usage. No tokens, no credits, no math.

  • Balance and spend per agent in dollars
  • A forecast of when your balance runs out
  • Prefer your own API key? Add it on top of your plan
Usage & Billing
Balance$24.60in dollars, no credits
Lasts about12 daysOn track
$0$25$50Sep 4Sep 11TodaySep 25Oct 2Oct 9Runs out ~Sep 30
OpenCode$0.92/day
OpenClaw$0.62/day
Hermes$0.51/day

Model costs go to your own provider account. Your plan still covers the agent in the cloud.

Update in one click

New OpenCode version? Make a backup and press update. If something breaks, roll back to your backup. Nothing to run in the terminal.

  • Make a backup before you update
  • Roll back to it if something breaks
  • No commands to run, no restarts to handle
Agents
OpenCodeNew version available
Create backupUpdate
Backup
Sep 28, 14:03Restore

Help with anything

Our team answers any question, from how to write your first prompt to automating your whole workflow.

  • Prompts, agents, models or billing: ask away
  • Real people from the Atomic Bot team
  • Priority support on every plan
# ask-the-team
M
maria_k

Which model should I pick for OpenCode?

Atomic Bot team TEAM

Start with GLM 5.3 for everyday coding. For a big refactor, try a Claude model: switch it in Settings in one click.

M
maria_k

Switched for the refactor, all tests pass. Thanks!

CHOOSE YOUR PLAN

Choose the best Cloud deployment plan for you.

Every plan includes AI usage: $5, $10 or $20 of your subscription goes to models each month

BASE

For set-and-forget users
$19/mo

Perfect to run your personal AI agent with instant access.


  • $5 of AI usage included every month
  • 1000+ connected apps
  • Fully encrypted and secure
  • 24/7 access to your agent
  • 100+ AI models
  • Priority support

System resources

vCPU cores 4
RAM 8 GB
SSD storage 80 GB
Subscribe to Base
Most popular

PRO

For daily agent users
$39/mo

More power for heavier workflows, more context and more speed.


  • $10 of AI usage included every month
  • 1000+ connected apps
  • Fully encrypted and secure
  • 24/7 access to your agent
  • 100+ AI models
  • Priority support

System resources

vCPU cores 8
RAM 16 GB
SSD storage 160 GB
Subscribe to Pro

MAX

For heavy users
$79/mo

Maximum capacity for demanding agents and intensive workloads.


  • $20 of AI usage included every month
  • 1000+ connected apps
  • Fully encrypted and secure
  • 24/7 access to your agent
  • 100+ AI models
  • Priority support

System resources

vCPU cores 16–64
RAM 32–128 GB
SSD storage 320–1280 GB
Subscribe to Max

FAQ

What is OpenCode?

OpenCode is an open-source AI coding agent. It reads your code, edits files and runs commands to get a task done. On Atomic Bot it runs in the cloud 24/7 and needs no install.

Where can I work with OpenCode?

In the chat on Atomic Bot, in Telegram, in the OpenCode dashboard and in its terminal interface (TUI). Open the dashboard or the TUI from your agent card in one click.

Can I run several OpenCode agents?

Yes. Run as many OpenCode agents as you need from one account, next to OpenClaw, Hermes and other agents. Each one keeps its own model and settings.

Can I use my own API key?

Yes. Open Settings, go to Inference and choose Use Your Own API Key. Or stay on Atomic Subscription and use 100+ models via OpenRouter without any key.

How much AI usage is included?

Base includes $5 a month, Pro $10 and Max $20. Usage & Billing shows every dollar you spend, and you can top up any time.

How do updates work?

When a new version is out, you update with one click. Make a backup first, and if something goes wrong, roll back to it.

Can I cancel my subscription anytime?

Yes. No lock-in, no commitments.