TEE Router

Chat

Uses only verified TEE routes, where the model runs inside attested confidential hardware. Billed at the chosen model's own price.

Price

Varies by chosen model

$0.12 to $4.80 per 1M input, $0.24 to $24.00 per 1M output

Privacy

TEE

Widest context

1M

Widest max output

1M

Quickstart

Use one model ID with your existing API key. Each request is billed at the chosen model's own price.

1. Get your API key

These examples use OpenAI-compatible access. Keys created in the dashboard have it on by default; a key without it is refused with compat_scope_required.

Create an API key

2. Make a request

Set the model to anonrouter/tee. The response's model field names the model that served. Do not send reasoning controls such as reasoning_effort, and do not pin a provider: the router picks the model and its route.

# Router: billed at the chosen model's own price. The response's model field names the model that served.
curl https://api.anonrouter.ai/v1/chat/completions -H "Authorization: Bearer $ANONROUTER_API_KEY" -H "Content-Type: application/json" -d '{
  "model": "anonrouter/tee",
  "max_tokens": 256,
  "messages": [
    {
      "role": "user",
      "content": "Hello"
    }
  ]
}'

Context and feature support depend on the chosen model. If no model in the router can serve your request, it is refused before anything is charged.

Pricing

Varies by chosen model. There is no router price: each request is billed at the chosen model's own price, settled from usage. Ranges below cover the router's current models, per 1M tokens.

Privacy levelInputOutput
TEE$0.12 to $4.80$0.24 to $24.00

Privacy

Uses only verified TEE routes: the model runs inside attested confidential hardware. It never falls back to a Private or Anonymous route. Requests are not end-to-end encrypted from your device; for that, choose an E2EE model.

Your workspace privacy boundary (Privacy page) always applies, and your workspace model pool (Routing page) narrows the router's members.

FAQ

Common questions about TEE Router.