Reka Research
Reka Research gives your app access to real-time web results in one call. Built for developers who need answers that are accurate, auditable, and grounded in evidence.
Fully compatible with OpenAI’s /chat/completions
format — including streaming and structured JSON. Swap in our model with minimal changes to your code.
How to use
- Use an official OpenAI SDK
- Change the following
- Update your base URL to point to Reka’s API
- Replace your API key with an Reka API key
- Update your model name to use
reka-flash-research
Here’s how to do this in Python, TypeScript, and Shell.
Response
Example response
Output
Each response includes:
- Chain-of-thought: Full trace of how the agent searched and reasoned (
message.reasoning_steps
) - Final reasoning: Internal rationale before the answer (
message.reasoning_content
) - Answer: Human-readable output with inline citations (
message.content
) - Citations: Start/end index, title, and URL for every source used
See Reasoning Steps for more information about the reasoning trace.
Citations
Citations are built in. Each one marks exactly where a source is used — with title, URL, and the span in the response it supports. Every url_citation
gives you precise auditability, down to the character index.