refactor · 14 files
✓ Diff ready
Long refactor
Hand off without losing context
Kick off a refactor at the desk. Close the laptop. The agent finishes in the Codespace and pings your phone when the diff is ready.
Live
$ codeam → workflow continuityYour AI workflow
doesn’t stop when you do.
Supervise, approve, and redirect Claude Code, Cursor, Copilot, Windsurf, Codex, and JetBrains AI from your phone — without ever opening a laptop.
This isn't coding on a tiny screen. It's the continuity layer that keeps your existing AI agents moving while you're away from the desk.
Async productivity · Workflow continuity · Team Spaces · Mobile Codespaces. Install once with codeam-cli, or grab the IDE plugin.
Free forever for 5 agent tasks a day. Upgrade only when your workflow can't afford to stop.
9:41
Claude Code
~/my-project · connected
Ask Claude anything...
Why nowAI agents went async.
AI agents went async.
Your workflow didn't.
Claude Code, Cursor, Copilot, and the agents already running in your IDE can now work for hours on their own — refactoring, testing, deploying. But the moment they need a decision, they stall. And you're stuck waiting at your desk for a "yes" or "no" that takes ten seconds.
The real problem isn't writing code from your phone. It's staying in the loop when your agent is working without you.
- Long agent runs stall waiting for approval.
- You can't leave your desk without losing the thread.
- Overnight runs pile up diffs you don't see until morning.
- Teammates can't jump into a session without restarting it.
- New ideas wait until you're back at a laptop.
The fix
One supervision layer for every agent you already use.
CodeAgent Mobile doesn't replace your agent, your IDE, or your workflow. It sits on top of what you already run — Claude Code, Cursor, Copilot, Windsurf, JetBrains AI, Codex — and gives you a mobile control surface to approve diffs, redirect prompts, and keep long runs on track from anywhere.
Pair once with a QR scan. From that point, your agent's session lives across every device you sign in on.
You approve and redirect from your phone — your IDE keeps running.
Long runs and overnight tasks queue diffs for you instead of stalling.
Your team can join a live session without restarting it.
Real workflows
See where the workflow takes over.
Three moments that don't work without a continuity layer — and feel obvious once you have one.
codespace
▶ Cloning repo
· Installing deps
· Starting Claude Code
· Ready ✓
Codespace deploy
Spin up a workspace from anywhere
Provision a GitHub Codespace from the app. Claude Code is preinstalled. Kick off the first task from the same screen.
E
M
A
Needs review
✓ Approved
Team review
Review your teammate's agent on the go
Get pinged when a teammate's shared session needs your eyes. Approve, comment, or redirect — without opening a laptop.
Workflow continuity
Three workflows your agents need when you’re not at the desk.
Supervise your IDE’s agents async, hand off live sessions to a teammate without restarting the thread, and provision cloud workspaces from any device. Your existing agent stack — kept moving while you’re away from the keyboard.
IDE supervision
Supervise the agents already in your IDE
Pair from your phone in one QR scan and stay in the loop with whichever agent your editor already runs — Claude Code, Copilot, Cursor, Windsurf, Vertex AI, JetBrains AI Assistant, and more. No new agent to learn, no environment migration.
- Official VS Code + JetBrains plugins
- QR pairing in seconds
- Voice or text — hands free for the coffee
- Works with the agents your IDE already installs
Team Spaces
Hand off live agent sessions to your team
Summon a paired agent into a team chat and share the live thread across developers. Anyone in the room can see what the agent is doing, redirect it, or take the keyboard — without restarting the session or losing context on the way.
- Group chat for developers
- Shared agent sessions across the room
- Live visibility into the agent’s turn
- Anyone on the team can redirect
- Integrated IDE for manual edits
Cloud workspaces
Spin up cloud workspaces from any device
New idea on the train? Pick a repo, branch, and machine size from your phone — Codespace boots, agent boots, CLI pairs, and the session is live by the time you get home. No 'I'll start it when I'm at my desk' delay.
- GitHub Codespaces today
- Repo, branch, and machine selection
- Claude Code preconfigured
- codeam-cli paired to your phone automatically
- GitLab, Bitbucket, and hosted VMs planned
Built for the moments you’re not at your desk
Six workflows your agents already need, served from wherever you are — async approvals, overnight runs, team handoffs, and mobile codespaces. The agent stack you already use, supervised continuously.
Approve agents from anywhere
Tap to approve diffs, redirect a tool call, or end a runaway turn — from a phone, between meetings, on a flight. Your IDE keeps running; you stay in the loop, async.
Supervise overnight runs
Long autonomous tasks grind while you sleep. Wake up to a queue of diffs ready to review — not a stalled cursor and an "are you still there?" prompt.
Catch wrong turns in seconds
Watch the agent's output stream live wherever you are. Spot a misstep, redirect with a tap, save the turn — instead of finding a 200-line wrong diff hours later.
Hand off without losing context
Share a live agent thread with a teammate — they join the running session instead of restarting from scratch. History, diffs, and pending approvals stay intact across the handoff.
Spin up workspaces from your phone
New idea on the train? Provision a sandboxed cloud workspace pre-paired with your linked agent in under a minute. No 'I'll start it when I'm at my desk' delay.
One layer, every agent you use
Claude Code, OpenAI Codex, Cursor, Windsurf Cascade, JetBrains AI, GitHub Copilot. Plug into whichever agent your IDE already runs — switching surfaces never costs you a workflow.
Recommended
The Fastest Way to Get Started
codeam-cli wraps Claude Code in your terminal — the same mobile control, no IDE required.
Terminal
1
Install the CLI
One global install from npm. Requires Node.js 18+ and Claude Code.
npm install -g codeam-cli2
Pair Your Phone
Run codeam pair, get a 6-character code, enter it in the CodeAgent Mobile app.
codeam pair3
Run & Control
Type codeam instead of claude. Your phone controls it in the background.
codeamWorks anywhere
SSH, Docker, remote servers — if you can run Node.js, you can run codeam
Identical mobile UI
Same CodeAgent Mobile app, same pairing flow, same real-time command relay
Multi-session
Pair multiple machines. Switch with codeam sessions switch
The Mobile App
Then Open the App
Scroll down to see the full experience — from login to sending your first prompt.
9:41
CodeAgent Mobile
email@example.com
••••••••
Sign In
Scan Pairing Code
Point camera at the QR from your terminal
A3F-K9P
Paired!
Connected to Edgar's MacBook Pro
Claude Code
online
Claude Code
online
Weekly quota: 23% used
Add dark mode toggle to settings page
I'll add a dark mode toggle. Let me update the settings component and create a theme context...
src/components/Settings.tsx
+ const [dark, setDark] = useState(true)
+ <Toggle checked={dark} />
Ask anything...
claude-sonnet-4-6
1
Sign in to the app
Download CodeAgent Mobile and create your account. Available on iOS and Android.
2
Scan the pairing code
Point your camera at the QR code from the terminal, or enter the 6-character code manually.
3
Instantly connected
Your phone pairs with the CLI session. Claude Code appears as an online agent ready for commands.
4
Control from anywhere
Send prompts, stream responses in real-time, manage files — all from your pocket.
Now from your phone — no laptop, no terminal
Cloud workspaces.Your laptop is optional.
Tap "New Codespace" in the app, pick a repo, and a fresh GitHub Codespace is provisioned, Claude Code is installed and authenticated, and your phone is paired — all without ever opening a laptop. Prefer the terminal? codeam deploy still works the same.
New
From the app — no terminal
Connect
Repo
Branch
Machine
Claude
Review
Deploy
Step 1 · Authorize GitHub
Connected as edgar-durand
~/your-project
$codeam deploy
✓Authorized with GitHub Codespaces
✓Workspace ready: friendly-telegram-…
✓Credentials extracted from macOS Keychain — no re-auth needed
✓Session running via PM2 — your phone stays paired
No laptop required
The agent runs on a Codespace, supervised by PM2. You can close your Mac, fly to another timezone, switch networks — the session stays paired to your phone.
Tap or type — your call
Launch from the mobile app in three taps, or run codeam deploy from any terminal. Both paths provision the workspace, install Claude, and pair your phone in ~60 seconds.
Pay only when coding
Tap "Stop session" in the app or run codeam deploy stop — the codespace pauses, compute hours stop. Resume with one command later, your branch and files exactly where you left them.
Async AI productivity
This isn't coding on glass. It's workflow continuity.
Your agent already does the writing. CodeAgent gives you the supervision layer on top — approvals, redirects, course-corrections, and team handoffs that follow you across devices instead of pinning you to one screen.
Kick off a long refactor from your desktop. Step into a meeting. Approve the diff from your phone. Resume on your laptop. The session never breaks.
Agent MemoryPick up exactly where
Pick up exactly where
your agent left off.
Across sessions and devices, your agent keeps a living task graph and persistent project memory — the decisions and conventions it has already learned. So it never re-learns your codebase, and never loses the thread.
READY
Persistent project memory
Decisions, conventions, and the why behind them stay with the agent. It reuses what you already settled instead of re-learning your codebase every new session.
IN PROGRESS
A live task graph
Ready, blocked, and in-progress work the agent tracks for itself. Close the laptop mid-refactor and the dependency graph picks up exactly where it left off.
SHARED
Team-shared context
Conventions the whole Team Space inherits — read-only, so everyone’s agent works from the same playbook without anyone re-explaining the project.
How it works
Pair once. From that point your agent’s session lives across every device you sign in on.
Pair an agent
QR-scan from the mobile app into your IDE plugin or CLI. The agent your editor already runs becomes the agent your phone supervises.
Stay in the loop
The agent keeps running on your machine; chunks stream live to your phone. Approve, redirect, or end a turn from anywhere.
Hand off without breaking
Walk away from the desk, pick the session back up on a laptop, hand it to a teammate. The thread keeps its history across every surface.
Every agent you already run
CodeAgent is the supervision layer — the agent stays whichever one your IDE or CLI already installs.
Cascade
Windsurf
Claude Code
Anthropic
JetBrains AI
Assistant
Junie
JetBrains
OpenAI Codex
OpenAI
GitHub Copilot
Cursor
Aider
aider.chat
CodeRabbit
AI Code Review
Gemini CLI
Google
Cascade
Windsurf
Claude Code
Anthropic
JetBrains AI
Assistant
Junie
JetBrains
OpenAI Codex
OpenAI
GitHub Copilot
Cursor
Aider
aider.chat
CodeRabbit
AI Code Review
Gemini CLI
Google
Plugins for your IDE
One install, one pair, and your phone is the supervision surface. Pick the plugin for your editor.
CodeAgent Mobile
for VS Code, Cursor & WindsurfVisual Studio Marketplace
Pair the editor with your phone. Stream Copilot Chat, Claude Code and Cascade responses to your phone, approve diffs, and drive prompts — even when you are away from the keyboard.
FreeGet from VS Code Marketplace
CodeAgent Mobile
for JetBrains IDEsJetBrains Marketplace
IntelliJ IDEA, WebStorm, PyCharm, GoLand, Rider, RubyMine, PhpStorm, CLion. Pair once, then drive Junie, JetBrains AI Assistant or any agent from your phone with a QR scan.
FreeGet from JetBrains Marketplace
Running an agent from a plain terminal? Skip the plugin — install codeam-cli instead.
Use cases
Picture your own day, solved
01
The Overnight Refactor
You kick off a 14-file refactor before bed. While you sleep, the agent works through it. In the morning, the diff is ready on your phone — review it over coffee, approve it before you've opened your laptop.
02
The Meeting Interruption
You're mid-task when a meeting starts. Instead of leaving the agent stuck waiting for input, you approve its next step from your phone under the table — and walk out of the meeting to a finished task.
03
The Commute Idea
You think of a fix on the train. Open the app, spin up a Codespace, and kick off the task — Claude Code is preconfigured and your phone is paired automatically. By the time you're home, there's a diff waiting.
04
The Team Handoff
A teammate's agent session needs a second opinion. Instead of a screen-share or a “can you check this later,” you join the live session from your phone, see exactly what the agent is doing, and weigh in immediately.
05
The Cross-Device Switch
You start a task on your work laptop, head out, and finish reviewing the diff on your personal laptop that evening. The session — history, diffs, pending approvals — follows you across devices.
Pricing
Start free. Upgrade when your workflow can’t afford to stop.
No credit card to start. Cancel anytime. 7-day money-back guarantee on your first Pro purchase.
Free
Try the full workflow. No credit card required.
$0forever
- 5 agent tasks per day
- 5 AI code reviews + 10 Smart Composer calls per day
- 5 @codeagent team mentions per day
- 1 Codespace deploy per day
- 2 paired devices · 1 linked agent
- Every agent your IDE already runs — Claude Code, Copilot, Cursor, Windsurf
- Team Spaces · voice prompts · QR pairing
Pro
For developers who ship from anywhere.
$9.99per month
or $99.99 / year — save ~17%
- Unlimited everything — no daily caps, no surprise walls mid-flow
- Pair every device you own — laptop + work laptop + phone + tablet
- Connect your team — share active sessions with up to 5 collaborators
- Always responsive — your agent never waits in a queue, even under load
- Direct line to engineering — replies under 24h on weekdays
- First access to new agents, integrations, and persistent cloud previews
Payments processed by Kelviq (our merchant of record). VAT / GST / sales tax calculated automatically based on your billing address. Read our refund policy and terms.
How CodeAgent Mobile compares
A continuity layer, not another mobile client
Several tools bring AI agents to mobile screens. CodeAgent Mobile is the workflow continuity layer underneath them — IDE pairing, async approvals, team handoffs, and cloud workspaces, supervising whichever agent your editor already runs.
Claude Code Remote Control
Focuses on
Continue local Claude Code sessions from phone, tablet, or browser.
Pricing: No free tier — requires Claude Pro ($20/mo) or Max ($100/mo).
What they do well
- Official Claude ecosystem
- Mobile and browser access
- Local Claude Code session continuation
- QR-style session handoff
Where CodeAgent Mobile goes further
- Works beyond Claude Code
- Supports IDE-based agents through VS Code and JetBrains plugins
- Includes voice prompts and QR pairing
- Adds Team Spaces for shared agent collaboration
- Can launch GitHub Codespaces from mobile
Omnara
Focuses on
Command center for coding agents like Claude Code and Codex.
Pricing: Free up to 10 agent sessions / month, Pro $9/mo unlimited, Enterprise contact sales.
What they do well
- Mobile, web, and desktop access
- Agent progress visibility
- Claude Code and Codex workflows
- Session monitoring and approvals
Where CodeAgent Mobile goes further
- More IDE-first with official VS Code and JetBrains plugins
- Controls existing IDE agents, not only CLI agents
- Includes Team Spaces for group AI sessions
- Integrated IDE in the mobile app for manual edits
- Supports mobile Codespace provisioning
Termly
Focuses on
Mobile bridge for terminal AI coding assistants.
Pricing: Free, no paid tier — optional "buy us a coffee" support.
What they do well
- CLI bridge to terminal AI tools
- QR setup and iOS / Android support
- Voice input
- Works with Claude Code, Copilot, OpenCode, Gemini
Where CodeAgent Mobile goes further
- Same QR pairing and voice prompts
- Adds official VS Code and JetBrains plugins
- Controls IDE-based agents, not only terminal sessions
- Adds Team Spaces and shared agent collaboration
- Adds Codespaces from mobile
SuperTurtle
Focuses on
Telegram-based autonomous coding agent.
Pricing: Commercial — pricing not public, sample comms reference ~$29/seat annual team plan.
What they do well
- Telegram control with text and voice messages
- Claude Code and Codex support
- Parallel workers / SubTurtles
- Runs on the user’s machine
Where CodeAgent Mobile goes further
- Dedicated native app instead of a Telegram bridge
- Official IDE plugin workflow
- Visual session and conversation management
- Team Spaces with shared agent sessions
- Integrated IDE and Codespaces support
Pricing at a glance
Public pricing snapshots from each vendor's site, May 2026. Check the linked source for the current rate.
| Product | Free tier | Paid tier |
|---|---|---|
| CodeAgent Mobile | 5 tasks/day · 2 devices | $9.99/mo · $99.99/yr (~17% off) |
| Claude Code Remote Control | None | Claude Pro $20/mo or Max $100/mo |
| Omnara | 10 agent sessions / month | Pro $9/mo · Enterprise contact sales |
| Termly | Free, no limits | n/a |
| SuperTurtle | Not public | Commercial — pricing not public |
Feature matrix
"Not primary" means the feature isn’t the tool’s main focus or wasn’t obvious from public docs at the time of writing.
| Feature | CodeAgent Mobile | Claude Remote Control | Omnara | Termly | SuperTurtle |
|---|---|---|---|---|---|
| Mobile app experience | Yes | Yes | Yes | Yes | Not primary |
| QR pairing | Yes | Yes | Not primary | Yes | Not primary |
| Voice prompts | Yes | Not primary | Not primary | Yes | Yes |
| Works with Claude Code | Yes | Yes | Yes | Yes | Yes |
| Works with multiple IDE agents | Yes | No | Not primary | Not primary | No |
| Official VS Code plugin | Yes | No | No | No | No |
| Official JetBrains plugin | Yes | No | No | No | No |
| Control existing IDE agents | Yes | No | Not primary | No | No |
| Team Spaces / group collaboration | Yes | No | Not primary | No | No |
| Shared agent session for a team | Yes | No | Not primary | No | No |
| Integrated IDE in app | Yes | No | No | No | No |
| Launch Codespaces from mobile | Yes | No | Not primary | No | No |
| Works without a laptop via cloud env | Yes | No | Not primary | No | No |
| Workflow-continuity positioning (not mobile coding replacement) | Yes | Not primary | Not primary | Not primary | Not primary |
Built for AI coding workflows
Not remote desktop. Built for the way AI agents actually work.
Remote desktop mirrors your laptop. CodeAgent Mobile connects directly to the way AI development actually works — prompts, agents, sessions, repositories, team spaces, and cloud environments.
| Capability | Remote Desktop | CodeAgent Mobile |
|---|---|---|
| Approve agents without unlocking the laptop | ||
| Pair an IDE with a QR scan | ||
| Voice-drive agents on the move | ||
| Resume long agent sessions from any device | ||
| Official VS Code + JetBrains plugins | ||
| Share live agent threads with a team | ||
| Spin up cloud workspaces from a phone | ||
| Browse + edit files without opening a laptop |
FAQ
Questions developers ask before signing up
Is this just coding on my phone?
No. CodeAgent Mobile isn't built for writing code on a small screen. It's a supervision layer — your agent does the work in your IDE or a Codespace, and your phone is how you approve, redirect, and stay in the loop without being at your desk.
Which AI agents does it work with?
Claude Code, GitHub Copilot, Cursor, Windsurf (Cascade), JetBrains AI Assistant, Junie, OpenAI Codex, Gemini CLI, Aider, and CodeRabbit — whichever agent your IDE or terminal already runs.
Do I need to install a new agent or change my workflow?
No. CodeAgent Mobile pairs with the agent you're already using. Install the VS Code or JetBrains plugin (or codeam-cli for terminal use), scan a QR code, and your phone becomes a supervision surface for that existing session.
What happens if I close my laptop mid-task?
If the agent is running locally, the session pauses with your machine. If you're running through a Codespace (via codeam deploy or the app), the agent keeps working in the cloud — you can close your laptop entirely and get notified when it's ready for review.
Can my team use this together?
Yes. Team Spaces let you share a live agent session with up to 5 collaborators (Pro). Everyone sees the same session, can redirect it, and the agent attributes prompts by name — no screen-share required.
Is there a free plan?
Yes. The Free plan includes 5 agent tasks per day, 2 paired devices, 1 linked agent, and access to every agent your IDE already runs. No credit card required.
What does Pro add?
Pro removes daily caps, lets you pair unlimited devices, adds Team Spaces for up to 5 collaborators, priority response times, and first access to new agents and integrations. $9.99/month or $99.99/year.
Does this work with terminal-only setups (SSH, Docker, remote servers)?
Yes. codeam-cli runs your agent in the terminal over the open ACP protocol — Claude Code, OpenAI Codex, Gemini, and other ACP-compatible agents — so it's not tied to one vendor. If you can run Node.js, you can run codeam, including over SSH or inside containers.
Is my code or data sent through CodeAgent's servers?
Your code stays where your agent runs — on your own machine or your own GitHub Codespace; we never clone your repository to our servers. CodeAgent Mobile acts as a relay: the prompts you send and the agent's streamed output (including diffs) pass through our backend to reach your phone, buffered only briefly and not retained. Team Space chat messages are stored so the conversation persists for your team. Your agent credentials are encrypted at rest.
Can I cancel anytime?
Yes. Pro is month-to-month or annual, cancel anytime, with a 7-day money-back guarantee on your first purchase.
Your agents are already working. Stay in the loop.
Install the CLI or IDE plugin, pair your phone in seconds, and keep your AI workflow moving — wherever you are.
Build in public
The workflow you need isn’t here yet?
The roadmap lives in GitHub Issues. If something you’d actually use is missing, open a request — we read every one, label it, and add the ones that fit the product direction to the next sprint.
01 · Open
File the issue in the public clients repo. Two or three sentences are enough — what workflow is broken without it.
02 · Label
Tag it enhancement so it lands in the feature backlog. We triage every week.
03 · Ship
If it’s aligned with the AI workflow continuity arc, it lands in a sprint. We reply publicly on the issue either way — no silent dismissals.
LIVE·CODEAGENT MOBILE·MOBILE-FIRST DEVELOPMENT·WORKSTATION-X9