Install
Initialize a client
Generate text
Configuration reference
Environment variables
Async usage
Move to /advanced-usage for retries, concurrency, and backoff patterns.
Documentation Index
Fetch the complete documentation index at: /llms.txt
Use this file to discover all available pages before exploring further.
Use the official ARIS Python client for generation, retries, and node routing.
pip install aris-sdk
from aris.client import Aris
client = Aris(
api_key="sk-aris-your-key",
registry_url="https://aris-api-production.up.railway.app/",
timeout=30,
)
response = client.generate(
prompt="Summarize the advantages of distributed inference.",
model="tinyllama",
max_tokens=256,
temperature=0.7,
)
print(response["output"])
| Setting | Type | Default | Description |
|---|---|---|---|
api_key | str | ARIS_API_KEY | Registry-issued API key. |
registry_url | str | http://localhost:8000 | Registry base URL. |
timeout | int | 30 | Request timeout in seconds. |
headers | dict[str, str] | {} | Extra headers for custom routing. |
| Variable | Description |
|---|---|
ARIS_API_KEY | API key for registry authentication. |
ARIS_REGISTRY_URL | Registry URL for discovery and handshake. |
import asyncio
from aris.client import AsyncAris
async def main() -> None:
client = AsyncAris(api_key="sk-aris-your-key")
result = await client.generate(prompt="Return a one-line health summary.")
print(result["output"])
asyncio.run(main())