Skip to main content
Answers to common questions about Anyware.

​
General

Anyware is a distributed Claude Code platform that runs sessions locally while enabling remote monitoring and control. Start Claude Code on your machine, then monitor and interact from any device via the web dashboard.Key features:
  • Claude Code runs locally with full codebase access
  • Real-time streaming to web dashboard
  • Switch between local and remote control
  • Support for multiple LLM models
  • SMS/WhatsApp notifications
AspectAnywareRegular Claude Code
Remote accessMonitor/control from any deviceLocal only
Model choiceUse GPT-4, Gemini, etc.Claude only
NotificationsSMS/WhatsApp alertsNone
Mode switchingLocal ↔ RemoteN/A
With Anyware, you get the same local Claude Code experience plus cloud connectivity.
Anyware: Claude Code runs locally on your machine. Your code stays on your computer. You get remote monitoring and control via the web dashboard.Catty: Claude Code runs in the cloud on isolated machines. Your workspace is uploaded to cloud sessions. Sessions persist when you disconnect.Choose Anyware if you want local execution with remote control. Choose Catty if you want fully cloud-hosted sessions.
Anyware is ideal for:
  • Developers who want to monitor Claude Code from their phone
  • Users who want to approve permissions while away from their desk
  • Anyone who wants to use models other than Claude
  • Teams that need visibility into AI coding sessions

​
Platform Support

Supported:
  • macOS (Intel and Apple Silicon)
  • Linux (x64 and ARM64)
Not yet supported:
  • Windows
Windows support is on the roadmap.
Node.js 20 or later is required to install the CLI via npm.
Not currently. The CLI is distributed via npm. You can use npx @diggerhq/anyware without a global install.

​
Sessions

No. Unlike Catty, Anyware sessions end when you exit Claude Code or Ctrl+C. Sessions are for real-time monitoring and control, not persistence.For persistent sessions, consider Catty or running Anyware on an always-on machine (see Running 24/7).
Yes! Run anyware in multiple terminal windows for different projects. Each session appears separately in the web dashboard.
Sessions are tied to the running Claude Code process. If you Ctrl+C or exit, the session ends. Start a new session with anyware.For Claude Code conversation persistence, use:
anyware --continue    # Continue last conversation
anyware --resume ID   # Resume specific session

​
Models

Anyware supports all models available through OpenRouter, including:
  • Claude (all versions)
  • GPT-4 and GPT-4o
  • Gemini
  • Llama
  • Mistral
  • And many more
See Custom Models for the full list.
No. Anyware routes requests through its proxy, which handles API access. Usage is billed through your Anyware subscription.
Reasons to try other models:
  • Cost: Some models are cheaper for simple tasks
  • Speed: Smaller models respond faster
  • Experimentation: Test different AI capabilities
  • Preference: Some tasks may work better with specific models

​
Security & Privacy

Yes. Claude Code runs locally on your machine. Your code is never uploaded anywhere. Only session events (terminal output, tool calls) are streamed to the cloud for the dashboard.
Locally at ~/.anyware/config.json. Credentials are only sent to the Anyware API for authentication and to the proxy for LLM requests.
No. Sessions are tied to your user account. Only you can view your sessions in the dashboard.
Yes. All connections use:
  • HTTPS for API calls
  • WSS (WebSocket Secure) for real-time streaming

​
Billing

Anyware bills based on token usage through the LLM proxy. See Pricing for current plans and limits.
No. Claude Code runs on your machine, so there’s no cloud compute cost. You only pay for LLM token usage.
When you reach your usage limit, you’ll be prompted to upgrade. You can upgrade immediately or wait for the next billing period.

​
Troubleshooting

Contact support with:
  • CLI version (anyware version)
  • OS and architecture
  • Full error message
  • Steps to reproduce