Skip to main content

Code Examples

These examples use the official OpenAI SDKs — they work with Schatzi AI by changing the base URL. Every snippet is copy-pasteable and ready to run.

Python (OpenAI SDK)​

Install the SDK:

pip install openai

Basic Chat Completion​

from openai import OpenAI

client = OpenAI(
    api_key="sk-your-api-key",
    base_url="https://backend.schatziai.ch/v1"
)

response = client.chat.completions.create(
    model="MODEL_ID",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Summarise the key points of FADP compliance."}
    ],
    max_tokens=500
)

print(response.choices[0].message.content)

Streaming​

stream = client.chat.completions.create(
    model="MODEL_ID",
    messages=[{"role": "user", "content": "Explain Swiss data sovereignty in 3 paragraphs."}],
    stream=True
)

for chunk in stream:
    content = chunk.choices[0].delta.content
    if content:
        print(content, end="", flush=True)

Embeddings​

response = client.embeddings.create(
    model="EMBEDDING_MODEL_ID",
    input="Swiss financial regulations"
)

vector = response.data[0].embedding
print(f"Dimensions: {len(vector)}")

JavaScript / TypeScript (OpenAI SDK)​

Install the SDK:

npm install openai

Basic Chat Completion​

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "sk-your-api-key",
  baseURL: "https://backend.schatziai.ch/v1",
});

const response = await client.chat.completions.create({
  model: "MODEL_ID",
  messages: [
    { role: "system", content: "You are a helpful assistant." },
    { role: "user", content: "Draft a client email in German and English." },
  ],
  max_tokens: 500,
});

console.log(response.choices[0].message.content);

Streaming​

const stream = await client.chat.completions.create({
  model: "MODEL_ID",
  messages: [{ role: "user", content: "List 5 benefits of Swiss data residency." }],
  stream: true,
});

for await (const chunk of stream) {
  const content = chunk.choices[0]?.delta?.content;
  if (content) process.stdout.write(content);
}

curl​

Basic Request​

curl https://backend.schatziai.ch/v1/chat/completions \
  -H "Authorization: Bearer sk-your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "MODEL_ID",
    "messages": [
      {"role": "user", "content": "What is FADP?"}
    ],
    "max_tokens": 200
  }'

Streaming​

curl https://backend.schatziai.ch/v1/chat/completions \
  -H "Authorization: Bearer sk-your-api-key" \
  -H "Content-Type: application/json" \
  -N \
  -d '{
    "model": "MODEL_ID",
    "messages": [{"role": "user", "content": "Count to 5"}],
    "stream": true
  }'

List Models​

curl https://backend.schatziai.ch/v1/models \
  -H "Authorization: Bearer sk-your-api-key"

Extended Parameters​

Reasoning Control (Python)​

For reasoning models, use reasoning_effort and max_completion_tokens to control the balance between speed and accuracy:

response = client.chat.completions.create(
    model="MODEL_ID",
    messages=[{"role": "user", "content": "Solve this step by step: what is 247 * 389?"}],
    extra_body={
        "reasoning_effort": "high",
        "max_completion_tokens": 16384,
    }
)

Set reasoning_effort to "low" for fast tool-calling workflows, or "high" for complex multi-step reasoning.

Structured JSON Output (Python)​

Request structured output using response_format:

response = client.chat.completions.create(
    model="MODEL_ID",
    messages=[{"role": "user", "content": "List 3 Swiss cities with population"}],
    extra_body={
        "response_format": {
            "type": "json_object"
        }
    }
)

import json
data = json.loads(response.choices[0].message.content)

Deterministic Output (curl)​

Use seed for reproducible results:

curl https://backend.schatziai.ch/v1/chat/completions \
  -H "Authorization: Bearer sk-your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "MODEL_ID",
    "messages": [{"role": "user", "content": "Generate a UUID"}],
    "seed": 42,
    "max_tokens": 50
  }'
info

Extended parameters (reasoning_effort, max_completion_tokens, response_format, seed, logprobs, top_logprobs, parallel_tool_calls) are passed through to the inference backend. Not all models support all parameters — if a model doesn't support a parameter, the API returns a 400 error. See Endpoint Reference for the full parameter list.

Migrating from OpenAI​

If you have existing code that uses the OpenAI API, switching to Schatzi AI requires only two changes:

SettingOpenAISchatzi AI
Base URLhttps://api.openai.com/v1https://backend.schatziai.ch/v1
API Keysk-... (OpenAI key)sk-... (Schatzi AI key from your dashboard)

Model IDs differ between platforms. Use GET /v1/models to discover available models.

export OPENAI_API_KEY="sk-your-schatzi-key"
export OPENAI_BASE_URL="https://backend.schatziai.ch/v1"

The OpenAI Python and JS SDKs read these environment variables automatically — no code changes needed beyond setting these variables.

Error Handling (Python)​

from openai import OpenAI, APIError, RateLimitError, AuthenticationError

client = OpenAI(
    api_key="sk-your-api-key",
    base_url="https://backend.schatziai.ch/v1"
)

try:
    response = client.chat.completions.create(
        model="MODEL_ID",
        messages=[{"role": "user", "content": "Hello"}]
    )
    print(response.choices[0].message.content)
except AuthenticationError:
    print("Invalid API key. Check your key in the dashboard.")
except RateLimitError:
    print("Usage limit reached. Check your billing period.")
except APIError as e:
    print(f"API error: {e.message}")

Tips​

  • Store your API key in an environment variable, not in source code.
  • Use streaming for long responses to improve perceived latency.
  • Set max_tokens to control response length and cost.
  • Use system messages to set consistent behaviour across requests.
  • See Model Comparison for choosing the right model.
  • See Monitoring Usage for tracking costs.