---
source_url: "https://docs.linkup.so/pages/documentation/get-started/introduction"
title: Introduction - Linkup API Documentation
mirrored_at: 2026-08-19T13:03:24.172Z
host: docs.linkup.so
cited_in_42a: true
mirror_canonical: "https://index.42a.ai/docs.linkup.so/pages/documentation/get-started/introduction"
---

> **Original source:** https://docs.linkup.so/pages/documentation/get-started/introduction

Get Started

The production-grade web search API for AI.

Learn how to build with Linkup by pasting this prompt to your agent:

Linkup is the production-grade web search API for AI. Teams at Legora, Polymarket, Cohere, McKinsey, and SNCF ship production AI on Linkup because four things have to be true:

-   **Accuracy.** [**Search**](https://docs.linkup.so/pages/documentation/endpoints/search/overview) ranks #1 among sub-second web search APIs on Verified SimpleQA with a 92% F-score. [**Research**](https://docs.linkup.so/pages/documentation/endpoints/research/overview) ranks #1 across the board on SealQA-0 with 61% accuracy. The [evaluation harness](https://github.com/LinkupPlatform/eval-simpleQA) is open source.
-   **Customization.** The index adapts to your domain. Define which sources to trust, which domains to prefer, which proprietary data to blend in.
-   **Security.** SOC 2 Type II, Zero Data Retention, GDPR, and Single Sign-On on every tier. Bring Your Own Cloud is available for the highest data sovereignty requirements: the full Linkup index runs inside your VPC and queries never leave your environment.
-   **Latency.** Sub-second on standard search. Fast enough for voice AI, sales copilots, and live research tools.

## Our endpoints

Linkup exposes five endpoints. Each has a distinct shape and is tuned for a different point on the latency/quality curve.

Endpoint

Shape

Latency

Mode

Use when

[**Search**](https://docs.linkup.so/pages/documentation/endpoints/search/overview)

Query → ranked URLs, sourced answer, or structured JSON

<1s–~30s

sync

Grounding an agent or LLM answer with retrieved web context

[**Fetch**](https://docs.linkup.so/pages/documentation/endpoints/fetch/overview)

URL → clean LLM-ready markdown

~1s

sync

Extracting the full content of a known page

[**Research**](https://docs.linkup.so/pages/documentation/endpoints/research/overview)

Query → detailed report or precise answer with inline citations

2–20 min

async

Multi-source synthesis, full reports, precise answers

[**Tasks**](https://docs.linkup.so/pages/documentation/endpoints/tasks/overview)

Batch of up to 100 Search/Fetch/Research calls → results for each

1s–20 min

async

Bulk, scheduled, or background workloads

[**Extract**](https://docs.linkup.so/pages/documentation/endpoints/extract/overview)

Seed URL + query → downloadable structured rows

Variable

async

Repeated records on a known page, such as products, jobs, or directories

## Where to next