Responses

The Responses API is suitable for clients, coding agents, and new application frameworks that use the OpenAI Responses protocol. It is commonly used when a unified input structure, tool calling, reasoning controls, or more complex response structures are needed.

Endpoint

POST /v1/responses

Full URL:

https://www.routeapi.ai/v1/responses

Request example

curl https://www.routeapi.ai/v1/responses \
  -H "Authorization: Bearer $ROUTEAPI_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o",
    "input": "Explain the main purpose of RouteAPI"
  }'

Common parameters

Field	Type	Required	Description
model	string	Yes	Model ID
input	string/array	Yes	Input content
instructions	string	No	System-level instructions
stream	boolean	No	Whether to stream output
tools	array	No	Tool definitions
tool_choice	string/object	No	Tool choice strategy
temperature	number	No	Sampling temperature
max_output_tokens	number	No	Maximum output token count

Differences from Chat Completions

Item	Chat Completions	Responses
Input structure	messages	input + instructions
Compatibility range	Most traditional OpenAI SDK compatible clients	More common in new protocol clients and coding agents
Tool calling	Supported, depends on the selected model	Supported, usually more expressive
Recommended scenarios	General chat and compatible access	New projects, coding agents, complex tasks

Integration suggestions

• If the client explicitly requires the OpenAI Responses protocol, use /v1/responses.
• If the client only supports traditional OpenAI Chat Completions, use /v1/chat/completions.
• Whether the same model supports Responses depends on RouteAPI model configuration and service capabilities.