If your stack already speaks OpenAI, you do not need a custom Scalattice SDK. Two common tools people already run are Open WebUI (self-hosted chat UI) and LiteLLM (Python SDK / proxy across many providers). Both can call Scalattice today.

Shared facts for both:

Base URL: https://api.scalattice.cloud/v1
Auth: Authorization: Bearer slt_...
Example model: qwen-3-8b (use any ID from GET /v1/models)
Get a key: Developers dashboard or scalattice setup

Open WebUI

Open WebUI connects to any OpenAI-compatible provider from Admin settings. No Scalattice plugin.

Admin Settings → Connections → OpenAI → Add Connection
URL: https://api.scalattice.cloud/v1
API Key: your slt_... key
Save. Models load from GET /v1/models. If one is missing, add its ID under Model IDs (Filter).
Select the model and chat. That traffic is billed like any other /v1/chat/completions call.

Full steps also live in Cloud docs: scalattice.cloud/docs/developers#open-webui.

LiteLLM

Use the built-in OpenAI-compatible path. Prefix the model with openai/ and pass Scalattice as api_base:

pip install litellm

import litellm

resp = litellm.completion(
    model="openai/qwen-3-8b",
    api_base="https://api.scalattice.cloud/v1",
    api_key="slt_...",
    messages=[{"role": "user", "content": "Say hello in one sentence."}],
)
print(resp.choices[0].message.content)

Proxy config sketch:

model_list:
  - model_name: qwen-scalattice
    litellm_params:
      model: openai/qwen-3-8b
      api_base: https://api.scalattice.cloud/v1
      api_key: os.environ/SCALATTICE_API_KEY

Cloud docs: scalattice.cloud/docs/developers#litellm. We are also landing a first-class scalattice/ provider entry upstream so you can drop api_base once it merges.

LiteLLM and Open WebUI on Scalattice

Open WebUI

LiteLLM

Links