GuideTool IntegrationsAPI ReferenceTerms & Policies
API ReferenceText SeriesOpenAI Native ProtocolPOSTResponses
Text SeriesPOSTResponses Routes
API Reference

Responses

POSThttps://api.tokenbay.com/v1/responses

OpenAI Responses API and Responses compact entry

Responses Routes

TokenBay provides the standard Responses path and the Responses compact path.

POST/v1/responses
POST/v1/responses/compact

Request

schema
modelstringRequired

An available model ID shown on the live Models page or in the console.

inputstring | array | objectRequired

Responses API input, including plain text, conversation input, or multimodal input.

input[].rolestringOptional

Input message role, such as user, assistant, or system.

input[].contentstring | array<object>Optional

Text or multimodal content blocks.

content[].typestringOptional

Content block type, such as input_text, input_image, or input_file.

content[].textstringOptional

Text content.

content[].image_urlstringOptional

Image URL or data URL.

content[].file_id / filenamestringOptional

File input identifier. Support depends on the upstream.

instructionsstringOptional

System-level or developer-level instructions.

streambooleanOptional

Set true to receive Responses SSE events.

max_output_tokensintegerOptional

Maximum output tokens.

temperature / top_pnumberOptional

Sampling parameters.

text / reasoningobjectOptional

Structured text output and reasoning configuration.

text.format.typestringOptional

Text format, such as text or json_schema.

text.format.schemaobjectOptional

Structured output schema.

reasoning.effortstringOptional

Reasoning effort.

reasoning.summarystringOptional

Whether to request a reasoning summary. Support depends on the model.

tools / tool_choicearray<object> | object | stringOptional

Tool calling configuration.

tools[].typestringOptional

Tool type, such as function, web_search, file_search, computer_use, and similar tools.

tools[].namestringOptional

Tool name.

tools[].parametersobjectOptional

Function tool parameter schema.

tool_choicestring | objectOptional

auto, none, required, or a specific tool.

parallel_tool_callsbooleanOptional

Whether parallel tool calls are allowed.

previous_response_idstringOptional

Continue from a previous Responses context.

includearray<string>Optional

Ask the upstream to include extra output fragments.

store / metadata / userboolean | object | stringOptional

Storage and client-side tracing fields. Do not include sensitive data.

Response

schema

Responses keep the Responses API shape. Streaming calls return Responses SSE events.

idstringOptional

Responses response ID.

objectstringOptional

Response object type.

modelstringOptional

Actual response model.

output[]array<object>Optional

Model output blocks, such as message, tool_call, or reasoning.

output[].typestringOptional

Output block type, such as message, function_call, reasoning, or tool_result.

output[].rolestringOptional

Message block role.

output[].content[]array<object>Optional

Output content blocks.

content[].typestringOptional

Content type, such as output_text, refusal, or annotation.

content[].textstringOptional

Output text.

output[].call_idstringOptional

Tool call ID.

output[].argumentsstring | objectOptional

Tool call arguments.

usageobjectOptional

Input, output, and total token usage when returned upstream.

usage.input_tokensintegerOptional

Input token count.

usage.output_tokensintegerOptional

Output token count.

usage.total_tokensintegerOptional

Total token count.

Related

Errors