Inference API

Create a chat completion

Generates the model's reply to a sequence of messages — with streaming, tool calls and structured output as with OpenAI.

POSThttps://api.futureway.ai/v1/chat/completions

Authentication

Header: Authorization: Bearer $FUTUREWAY_API_KEY
Scope: inferenceInference API: chat completions, embeddings, model list.
Key kinds: personal · service
Format: OpenAI-compatible
Streaming: yes — stream: true returns server-sent events

Details on keys and scopes under Authentication.

Parameters

Name	Type	Required	Description
model	string	yes	Model alias, e.g. futureway-smart.
messages	array	yes	The conversation in order; at least one message.
messages[].role	stringsystem \| user \| assistant \| tool	yes	Who is speaking.
messages[].content	string	yes	Text of the message; images as content parts as with OpenAI, if the model supports images.
stream	boolean	no	Answer as server-sent events as tokens are generated.
temperature	number	no	Randomness 0–2; the model's default when omitted.
max_tokens	integer	no	Upper bound on the answer length.
tools	array	no	Tool definitions in the OpenAI function format.
tool_choice	string	no	auto, none or a specific tool.
response_format	object	no	Structured output, e.g. { "type": "json_object" }.

Example request

curl https://api.futureway.ai/v1/chat/completions \
  -H "Authorization: Bearer $FUTUREWAY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "futureway-smart",
  "messages": [
    {
      "role": "user",
      "content": "Was ist die Hauptstadt von Frankreich?"
    }
  ]
}'

Response

200

{
  "id": "chatcmpl-8f3c…",
  "object": "chat.completion",
  "created": 1757750400,
  "model": "futureway-smart",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Die Hauptstadt von Frankreich ist Paris."
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 14,
    "completion_tokens": 9,
    "total_tokens": 23
  }
}

Errors

Status	Code	Description
401	unauthorized	No, invalid or revoked API key.
400	invalid_request_error	Invalid request (missing required fields, unknown parameter, body too large).
404	model_not_found	Unknown model id, or a model this key may not address.
429	rate_limit_error	Too many requests in the window; retry-after names the wait.
429	budget_exceeded	The organisation's usage quota is exhausted — no retry helps until the window resets.
503	service_unavailable	The model is temporarily unreachable — retry with backoff.

Formats and handling under Errors.