# CARouter

> CARouter is an OpenAI-compatible API gateway in front of AI providers that run their inference on Canadian soil. One base URL, one key, and the choice of which companies - in which province, under which ownership - may see your prompts. Billed in CAD.

## How to call it

```python
from openai import OpenAI

client = OpenAI(
    base_url="https://carouter.ai/v1",
    api_key="car_...",  # from https://carouter.ai/dashboard/keys
)
response = client.chat.completions.create(
    model="command-r7b-12-2024",
    messages=[{"role": "user", "content": "Bonjour"}],
)
```

Endpoints: `POST /v1/chat/completions`, `POST /v1/completions`, `POST /v1/responses`, `POST /v1/embeddings`, `GET /v1/models`. Streaming with `"stream": true` on everything except embeddings. Errors use OpenAI's shape: a top-level `{"error": {...}}`.

## Documentation

- [CARouter for language models](https://carouter.ai/llms-full.txt): every endpoint, every model, every price and every error code, in one plain-text fetch. Read this one.
- [Documentation](https://carouter.ai/docs): the same reference as a web page, with live prices per provider.
- [Model catalog JSON](https://carouter.ai/api/public/models): the catalog this file is generated from. Takes `?plan=free|pro|scale`.
- [Provider network](https://carouter.ai/#providers): who serves traffic, where, and who owns them.

## Models (14 routable, 16 announced)

Prices are list prices on the Free plan (15% over the provider's base rate), in CAD per million tokens.

- `command-a-plus-05-2026`: 125K context; Cohere commercial; vision; tools; 3.9445/15.755 CAD per M tokens in/out; routable now
- `command-a-03-2025`: 111B; 250K context; Cohere commercial; tools; 3.9445/15.755 CAD per M tokens in/out; routable now
- `command-a-reasoning-08-2025`: 250K context; Cohere commercial; reasoning; tools; 3.9445/15.755 CAD per M tokens in/out; routable now
- `command-a-vision-07-2025`: 125K context; Cohere commercial; vision; tools; 3.9445/15.755 CAD per M tokens in/out; routable now
- `command-a-translate-08-2025`: 8K context; Cohere commercial; 3.9445/15.755 CAD per M tokens in/out; routable now
- `command-r-plus-08-2024`: 125K context; Cohere commercial; tools; 3.9445/15.755 CAD per M tokens in/out; routable now
- `command-r-08-2024`: 125K context; Cohere commercial; tools; 0.2415/0.943 CAD per M tokens in/out; routable now
- `command-r7b-12-2024`: 7B; 125K context; Cohere commercial; tools; 0.0575/0.2415 CAD per M tokens in/out; routable now
- `c4ai-aya-expanse-32b`: 32B; 125K context; CC-BY-NC 4.0; 1.104/1.104 CAD per M tokens in/out; routable now
- `c4ai-aya-vision-32b`: 32B; 16K context; CC-BY-NC 4.0; vision; 1.104/1.104 CAD per M tokens in/out; routable now
- `tiny-aya-global`: 8K context; CC-BY-NC 4.0; 0.161/0.161 CAD per M tokens in/out; routable now
- `gpt-oss-20b`: 21B; 128K context; Apache 2.0; reasoning; tools; 0.0741/0.2775 CAD per M tokens in/out; announced, not routable yet
- `qwen3.6-27b`: 27B; 256K context; Apache 2.0; reasoning; vision; tools; 0.7401/4.9956 CAD per M tokens in/out; announced, not routable yet
- `qwen3.5-397b-a17b`: 397B; 256K context; Apache 2.0; reasoning; vision; tools; 1.1101/6.6608 CAD per M tokens in/out; announced, not routable yet
- `qwen3.5-9b`: 9.7B; 256K context; Apache 2.0; reasoning; vision; tools; 0.185/0.2775 CAD per M tokens in/out; announced, not routable yet
- `qwen2.5-vl-72b-instruct`: 72B; 32K context; Qwen; vision; 1.6837/1.6837 CAD per M tokens in/out; announced, not routable yet
- `glm-5.2`: 256K context; MIT; reasoning; tools; 3.3304/10.1763 CAD per M tokens in/out; announced, not routable yet
- `deepseek-v4-flash-0731`: 256K context; MIT; reasoning; tools; 0.7401/1.4802 CAD per M tokens in/out; announced, not routable yet
- `qwen3.6-35b-a3b`: 35B; 256K context; Apache 2.0; reasoning; vision; tools; 0.4625/2.7754 CAD per M tokens in/out; announced, not routable yet
- `qwen3-235b-a22b-instruct-2507`: 235B; 244K context; Apache 2.0; tools; 1.3877/4.163 CAD per M tokens in/out; announced, not routable yet
- `qwen3-coder-30b-a3b-instruct`: 30B; 128K context; Apache 2.0; tools; 0.3701/1.4802 CAD per M tokens in/out; announced, not routable yet
- `gemma-4-26b-a4b-it`: 26B; 256K context; Apache 2.0; reasoning; vision; tools; 0.4625/0.9252 CAD per M tokens in/out; announced, not routable yet
- `mistral-medium-3.5-128b`: 128B; 176K context; Modified MIT; reasoning; vision; tools; 2.7754/13.8768 CAD per M tokens in/out; announced, not routable yet
- `mistral-small-3.2-24b-instruct-2506`: 24B; 128K context; Apache 2.0; vision; tools; 0.2775/0.6476 CAD per M tokens in/out; announced, not routable yet
- `pixtral-12b-2409`: 12B; 128K context; Apache 2.0; vision; tools; 0.3701/0.3701 CAD per M tokens in/out; announced, not routable yet
- `embed-v4.0`: 125K context; Cohere commercial; vision; 0.184/0.00 CAD per M tokens in/out; routable now
- `embed-multilingual-v3.0`: 0K context; Cohere commercial; 0.161/0.00 CAD per M tokens in/out; routable now
- `embed-english-v3.0`: 0K context; Cohere commercial; 0.161/0.00 CAD per M tokens in/out; routable now
- `qwen3-embedding-8b`: 7.6B; 32K context; Apache 2.0; 0.185/0.00 CAD per M tokens in/out; announced, not routable yet
- `bge-multilingual-gemma2`: 8K context; Gemma; 0.0185/0.00 CAD per M tokens in/out; announced, not routable yet

## Providers (1 live, 14 announced)

A provider that is not both Canadian-owned and Canadian-hosted is switched OFF for new accounts and stays off until the account holder turns it on.

- Cohere (Iowa (GCP us-central1), United States; Canadian-owned; zero data retention; live; off by default)
- Denvr Dataworks (Calgary, AB; Canadian-owned; zero data retention; announced; on by default)
- Bell AI Fabric (Kamloops, BC; Canadian-owned; zero data retention; announced; on by default)
- TELUS Sovereign AI (Rimouski, QC; Canadian-owned; zero data retention; announced; on by default)
- Hypertec Cloud (Montreal, QC; Canadian-owned; zero data retention; announced; on by default)
- QScale (Levis, QC; Canadian-owned; zero data retention; announced; on by default)
- eStruxture AI Cloud (Montreal, QC; Canadian-owned; zero data retention; announced; on by default)
- ThinkOn (Toronto, ON; Canadian-owned; zero data retention; announced; on by default)
- Aptum (Toronto, ON; Canadian-owned; zero data retention; announced; on by default)
- Distributive (Kingston, ON; Canadian-owned; zero data retention; announced; on by default)
- Amazon Bedrock (ca-central-1) (Montreal, QC; foreign-owned; zero data retention; announced; off by default)
- Azure OpenAI (Canada East) (Quebec City, QC; foreign-owned; zero data retention; announced; off by default)
- Google Vertex AI (northamerica-northeast1) (Montreal, QC; foreign-owned; zero data retention; announced; off by default)
- Scaleway (Paris, France; foreign-owned; zero data retention; announced; off by default)
- OVHcloud Canada (FR endpoints) (Gravelines, France; foreign-owned; zero data retention; announced; off by default)

## Notes

- CARouter is in alpha. The gateway, metering and billing are real; the provider network is still being signed up, which is why models are marked routable or announced above.
- Prompts and completions are not stored. Request metadata (model, provider, token counts, latency, cost) is, because billing and usage history are made of it.
- Generated 2026-08-14 from the live catalog.