OpenAI-style API

Base address: https://dev.typno.ai/v1. Authenticate with Authorization: Bearer tpk-....

Chat completions

POST /v1/chat/completions takes the same body as OpenAI: model, messages, and optionally stream, max_tokens, temperature, top_p, stop, tools, tool_choice and response_format.

curl https://dev.typno.ai/v1/chat/completions \
  -H "Authorization: Bearer $TYPNO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "mistral4-119b",
    "messages": [
      {"role": "system", "content": "You answer briefly."},
      {"role": "user", "content": "What is the capital of the Netherlands?"}
    ]
  }'

Streaming

Set "stream": true to get server-sent events with the answer as it is written. Add "stream_options": {"include_usage": true} for a final chunk with the token usage.

stream = client.chat.completions.create(
    model="mistral4-119b",
    messages=[{"role": "user", "content": "Write a haiku about rain."}],
    stream=True,
)
for chunk in stream:
    print(chunk.choices[0].delta.content or "", end="")

Tools

Pass functions in tools; the model answers with tool_calls, you run them and send the results back as role: "tool" messages.

{
  "model": "glm5.3-754b",
  "messages": [{"role": "user", "content": "What is the weather in Utrecht?"}],
  "tools": [{
    "type": "function",
    "function": {
      "name": "get_weather",
      "description": "Current weather for a city",
      "parameters": {"type": "object", "properties": {"city": {"type": "string"}}, "required": ["city"]}
    }
  }]
}

Images in a message

Models with vision take images as image_url parts, as a public address or a data: URL.

{"role": "user", "content": [
  {"type": "text", "text": "What is in this picture?"},
  {"type": "image_url", "image_url": {"url": "data:image/png;base64,..."}}
]}

Not included

There is no web search or file upload in the API; those exist in the chat only. Embeddings, audio and fine-tuning are not offered.