Introducing ModelTunnel AI Native Cloud: Better frontier models, faster runs, lower costs →

App and Service Availability

Clients

ClientStatus
Web — ChatAvailable on all modern desktop and mobile browsers
Android appAvailable as an APK from the APP menu (Android 8.0 or later)
iOS appComing soon — use the web app and Add to Home Screen in the meantime
ModelTunnel Claw (desktop agent)macOS 11 or later; Windows coming soon
ModelTunnel Code (terminal coding agent)macOS, Linux and Windows (Python 3.10 – 3.13)

API

The API at https://modeltunnel.ai/v1 is served globally. Requests are routed to the upstream provider that hosts each model; latency depends on the model and your region.

  • Streaming — long responses should use "stream": true; if the client disconnects, only the tokens generated so far are billed.
  • Data retention — API prompts and completions are not stored by ModelTunnel. Only usage metadata (model, token counts, cost) is kept for billing. Chat conversations are saved to your account until you delete them.

Status and support

If a model returns 5xx errors, the upstream provider is usually degraded — retry with backoff or switch to another model (or use auto). For account and billing questions, email support@modeltunnel.ai.

App and Service Availability · ModelTunnel