Put your machines to work

Ask your AI client to manage a home server, run a training job or investigate a production issue. Explore examples for the hardware you already own.

Home & self-hosted
Home server & Raspberry Pi, from anywhere
Restart a stuck service, free up disk, or check why the Pi is hot — from your phone, without port forwarding or a VPN.
Replaces VNC / port forwarding
Run your self-hosted home lab by chat
Pull and restart Docker containers, update your media server, or check Pi-hole by asking instead of opening SSH.
Remote access for AI agents
Manage a NAS & backups remotely
Check SMART health, restart a stuck share, or confirm last night's backup without being home.
Headless, no port forwarding
Remote tech support for someone else's machine
Fix a relative's home server or a friend's VPS from your own chat: logs, services, packages.
No screen-share call needed
Developers & cloud
Operate cloud VMs without exposing SSH
Tail logs, restart a service, or run a migration on a production VM with no public port 22 and no bastion.
No SSH port · works behind NAT
Drive a remote GPU / ML training box
Operate a headless rig from your laptop: watch nvidia-smi, fix a CUDA error, free a stuck card, or grab the loss curve.
Headless, no display needed
Run ML training that survives the chat
Ask Claude to start the training, then close everything. The run keeps going on your own GPU — come back tomorrow and ask how it went.
Detached jobs · your own hardware

Claude Code · Codex · ChatGPT harness

Not a replacement for those tools — the remote-machine layer they plug into over MCP.

Grok CLI / grok build harness

Keep Grok. Give it your machines over MCP — remote shell, jobs, and file transfer. Not a replacement.

Cursor MCP harness

Keep Cursor. Give it your machines over MCP — remote shell, jobs, and file transfer. Not a replacement.

Codex CLI harness

Keep Codex. Give it your machines over MCP — remote shell, jobs, and file transfer. Not a replacement.

ChatGPT / OpenAI harness

Keep ChatGPT, GPT, and submodels like Astra (gpt-6-astra). Give them your machines over MCP. Not a replacement.

Run coding agents on a remote machine
Let Claude Code or Codex work in a repo on your own box as a detached job — start it, close the chat, read the diff tomorrow.
Agent-to-agent · survives the chat
Debug CI runners & build servers live
Reproduce a flaky build on the actual runner, inspect the workspace, and clear caches interactively.
Your real runner, not a sandbox
Unattended automation & scheduled AI jobs
Let a cron job, CI step, or one AI agent calling claude -p run on a box with no operator present — a non-root, allowlisted service token instead of an interactive code.
Sandboxed · non-root · revoke anytime
Move files to and from a remote machine
On Pro, pull an artifact or push a config with the same MCP and REST workflow as commands and jobs.
File transfer · Pro · up to 100 MiB per file
Enterprise & fleets
Bulk-update & patch a whole fleet
Apply security updates across Linux boxes, then verify versions from one conversation.
Fleet patching by chat
Install software across many machines
Roll out an agent, CLI, or config change to dozens of servers and confirm it landed everywhere.
Unattended, no per-machine login
On-call incident response from your phone
When a service is down, triage logs, restart it, and confirm recovery by chatting from anywhere.
No laptop, no VPN
Manage IoT & edge devices remotely
Update an app, restart a sensor service, or read logs on Pis and gateways, even behind NAT or cellular.
Outbound-only, fleet-wide

One setup for every example

Install the agent and connect your AI client. The machine connects outbound, so you do not need inbound ports or a VPN. Follow the setup guide.

To collect a report, checkpoint or rendered image, or send a config file, add file transfer with Pro.

Pick a machine and try it

Install the agent on one Linux box, or run the desktop app on a Mac or PC, and ask your AI client to do something on it. Fresh codes work for the first hour without an account; sign in for ongoing access.