Reka AI Documentation

1 curl -X POST https://api.reka.ai/v1/chat/completions \ 2 -H "Authorization: Bearer <token>" \ 3 -H "Content-Type: application/json" \ 4 -d '{ 5 "messages": [ 6 { 7 "role": "user", 8 "content": "foo" 9 } 10 ], 11 "model": "reka-flash-research" 12 }'

1 { 2 "id": "foo", 3 "choices": [ 4 { 5 "finish_reason": "stop", 6 "index": 42, 7 "message": { 8 "role": "assistant", 9 "reasoning_steps": [ 10 { 11 "role": "assistant", 12 "reasoning_content": "foo", 13 "tool_calls": [ 14 { 15 "id": "foo", 16 "name": "search_web", 17 "args": { 18 "urls": [ 19 "foo" 20 ], 21 "task": "foo" 22 } 23 } 24 ], 25 "content": { 26 "tool_name": "search_web", 27 "tool_output": [ 28 { 29 "url": "foo", 30 "title": "foo", 31 "snippet": "foo", 32 "age": "foo" 33 } 34 ], 35 "tool_error": "foo" 36 }, 37 "tool_call_id": "foo" 38 } 39 ], 40 "reasoning_content": "foo", 41 "content": "foo", 42 "annotations": [ 43 { 44 "type": "url_citation", 45 "url_citation": { 46 "end_index": 42, 47 "start_index": 42, 48 "url": "foo", 49 "title": "foo" 50 } 51 } 52 ] 53 } 54 } 55 ], 56 "created": 42, 57 "model": "foo", 58 "usage": { 59 "completion_tokens": 0, 60 "prompt_tokens": 0, 61 "total_tokens": 0 62 } 63 }

Request

This endpoint expects an object.

messageslist of objectsRequired

A list of messages comprising the conversation so far.

modelstringRequired

The model name used to generate the response. Currently, only "reka-flash-research" is supported for this endpoint.

streambooleanOptionalDefaults to false

Determines whether to stream the response incrementally. If set to true, the agent will return a stream of response chunks. If false, the response will be returned as a single object.

researchobjectOptional

response_formatobjectOptional

An object specifying the format that the agent must output. Setting this to { "type": "json_schema", "json_schema": {...} } enables Structured Outputs, which ensures the agent returns output that matches the supplied JSON schema. If not specified, the agent will return a plain text response.

Response

CreateChatCompletionResponseobject

A chat completion response from the agent, containing the generated message and other metadata. This is used when stream is set to false in the request.

CreateChatCompletionStreamResponseobject

A streamed response from the agent, containing a chunk of the chat completion. This is used when stream is set to true in the request.

Chat Completions

Headers

Request

Response