LiteLLM Integration
Integrate Thesean with LiteLLM to call supported Ship models through the Anthropic-compatible Messages API.
Ensure you have your Thesean API key from the Thesean Dashboard before continuing.
Overview
LiteLLM provides:
- A proxy server for centralized access control, spend tracking, and rate limiting
- A Python SDK for load balancing and cost tracking
LiteLLM Python SDK
Installation
uv add litellm
Basic Usage
from litellm import completion
import os
os.environ["THESEAN_API_KEY"] = "your-thesean-api-key"
response = completion(
model="anthropic/ship-like/claude-opus-4-8",
api_base="https://api.thesean.ai",
api_key=os.environ["THESEAN_API_KEY"],
max_tokens=1024,
messages=[{"role": "user", "content": "Hello!"}],
)
print(response.choices[0].message.content)
Ship Models
Use LiteLLM's anthropic/ provider prefix before the Ship model name:
LiteLLM consumes the leading anthropic/ prefix to select its Anthropic transport. Thesean receives
the remaining ship-like/claude-opus-4-8 model ID.
response = completion(
model="anthropic/ship-like/claude-opus-4-8",
api_base="https://api.thesean.ai",
api_key=os.environ["THESEAN_API_KEY"],
max_tokens=1024,
messages=[{"role": "user", "content": "Hello!"}],
)
LiteLLM Proxy Server
Create a config.yaml file:
model_list:
- model_name: ship-opus
litellm_params:
model: anthropic/ship-like/claude-opus-4-8
api_base: https://api.thesean.ai
api_key: "os.environ/THESEAN_API_KEY"
general_settings:
master_key: sk-1234
Install and start the proxy:
uv tool install "litellm[proxy]"
litellm --config config.yaml
Use the bare origin for api_base. LiteLLM's Anthropic transport appends /v1/messages
automatically.
Send a Messages request through the proxy:
curl http://0.0.0.0:4000/v1/messages \
-H "Content-Type: application/json" \
-H "Authorization: Bearer sk-1234" \
-d '{
"model": "ship-opus",
"max_tokens": 1024,
"messages": [
{
"role": "user",
"content": "Hello!"
}
]
}'