Responses API

# High-quality web research, in seconds

## Build fast and accurate web research experiences into your product.

Features

Fast, predictable, drop-in

High quality, low latency

Responses returns simple questions in just a couple seconds and can do deep research inside a minute. Perfect for conversational agents and user-facing UIs.

Get Started[Get Started]

Flat, predictable cost

Each Responses API call costs the same, no matter how many tokens Parallel consumes under the hood. Scale usage confidently from 1 call to 1 million calls.

Get Started[Get Started]

OpenAI compatible

Use Parallel's Responses API through OpenAI's Python or TypeScript SDKs. Switching to Parallel is a three-line diff.

Get Started[Get Started]

  from openai import OpenAI - client = OpenAI(api_key=OPENAI_API_KEY)+ client = OpenAI(api_key=PARALLEL_API_KEY,+                 base_url="https://api.parallel.ai/v1")   response = client.responses.create(-     model="gpt-5.4-mini",+     model="parallel",      input="What did the SEC change for broker-dealers?",      reasoning={"effort": "low"},-     tools=[{"type": "web_search"}],  )

Use Cases