# LiteLLM setup

LiteLLM gives many providers one interface, as a Python library or as a proxy server. Claudech plugs in as an OpenAI-compatible backend.

## Before you start

- **A Claudech API key** from [API keys](https://claudech.com/api-keys), stored as `CLAUDECH_API_KEY`. API access unlocks after your first purchase.

## Python SDK

```python
# pip install litellm
import os
from litellm import completion

response = completion(
    model="openai/claude-opus-5-5",
    api_base="https://api.claudech.com/v1",
    api_key=os.environ["CLAUDECH_API_KEY"],
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)
```

The `openai/` prefix selects LiteLLM's OpenAI-compatible route; the part after it is the Claudech model id.

## Proxy server

Add Claudech models to the proxy's `config.yaml`:

```yaml
model_list:
  - model_name: claude-opus-5-5
    litellm_params:
      model: openai/claude-opus-5-5
      api_base: https://api.claudech.com/v1
      api_key: os.environ/CLAUDECH_API_KEY
  - model_name: claude-sonnet-5
    litellm_params:
      model: openai/claude-sonnet-5
      api_base: https://api.claudech.com/v1
      api_key: os.environ/CLAUDECH_API_KEY
```

Start it with `litellm --config config.yaml`. Clients of the proxy then request `claude-opus-5-5` or `claude-sonnet-5` by `model_name`.

## Things to know

- LiteLLM's own cost tracking does not know Claudech prices. The exact cost of every request is in the `usage.claud_cost_usd` field and on your [Usage](https://claudech.com/usage) page.

See [all integrations](https://claudech.com/docs/integrations) for other tools. Claudech is an independent service and is not affiliated with the makers of LiteLLM.

---

Source: https://claudech.com/docs/integrations/litellm