One API for
every web data need.

Scrape, search, crawl, extract — one unified platform. Production-ready in 3 lines. Start free with 10,000 requests / month.

⌘K
Try
POST /v1/fetch
complete312ms
02Playground

See the full pipeline.

Pick a query and watch ZapFetch search, fetch, and structure in real time. Every sample is a real response shape.

Input
search
Query
latest OpenAI research papers
Mode
searchfetchstructure
Schema
{ results: Array, meta: Object }
Status
Search
Fetch
Structure
Discovered URLs
0/5
Oopenai.com142ms
Deliberative Alignment: Reasoning Enables Safer LLMs
https://openai.com/research/deliberative-alignment
Oopenai.com168ms
GPT-4o System Card
https://openai.com/index/gpt-4o-system-card/
Oopenai.com201ms
Weak-to-strong generalization
https://openai.com/research/weak-to-strong-generalization
Aarxiv.org233ms
Improving alignment via self-play — arXiv
https://arxiv.org/abs/2407.21787
Oopenai.com267ms
Sora system card
https://openai.com/research/sora-system-card
Structured JSON
awaiting query...
Search142ms
Fetch487ms
Structure89ms
Tokens1024
Cost$0.0012
03How it works

Three primitives.
One pipeline.

ZapFetch composes search → fetch → structure into a single streaming call, so your agent doesn't have to orchestrate three providers.

01Discovery
Search
Query across indexes — Exa, Tavily, Brave, Bing — in parallel. Rank by freshness, rerank by relevance.
02Rendering
Fetch
Pull pages at sub-second p50. JS rendering when needed. Proxy rotation, retries, and robots awareness built in.
03Extraction
Structure
Schema-guided extraction into typed JSON. Zod / Pydantic schemas supported. Streaming for long bodies.
input"latest OpenAI papers"
output{ results: [...], meta: {...} }
elapsed487ms
04Built for scale

Everything you need. Nothingyou don't.

412ms
p50
Sub-second scrape
Global edge, proxy warm-pool, and aggressive caching. Your agent doesn't wait.
Schema-first extraction
Pass a Zod or Pydantic schema; get back validated JSON. No regex, no cleanup.
schema: z.object({ ... })
Multi-source search
Parallel query across Exa, Tavily, Brave, Bing — dedup + rerank in one call.
exatavilybravebing
Full JS rendering
SPA and dynamic content just work. We run headless Chromium when static fails.
Batch-aware
Queue up to 1,000 URLs per request. Poll, or stream status events.
42 / 60 done10 pending
Server-sent streaming
Stream extracted fields as soon as they're parsed. No waiting on last byte.
data: {"title":"..."}
data: {"author":"..."}
data: {"posts":[...]}
05Compare

Not a search API.
Not a scraper.
The layer between.

ZapFetch sits where your LLM needs it: one call, three primitives, typed output. Existing tools force you to stitch them yourself.

Capability
ZapFetch
search + fetch + structure
Exa
neural search
Firecrawl
web scraping
Scope
Search + Fetch + Structure
Search only
Fetch only
Primary use case
AI agents, pipelines
Neural search API
Web scraping
p50 latency
412ms
~700ms
~900ms
Schema-guided extract
Streaming output
N/A
limited
JS rendering
Multi-source search
Batch up to 1,000 URLs
yes
async jobs
Typical cost / 1k req
$2.40
$5.00
$3.00
~/architecture
[search API][zapfetch][LLM / agent]
06Developer-first

$ npm install @zapfetch/sdk

Typed SDKs, a REST API, and docs that don't waste your time. Production in 3 lines.

import { Zapfetch } from '@zapfetch/sdk';
import { z } from 'zod';

const client = new Zapfetch({ apiKey: process.env.ZF_KEY });

// One call: search, fetch, structure
const { data } = await client.fetch({
  query: 'latest OpenAI research papers',
  schema: z.object({
    results: z.array(z.object({
      title: z.string(),
      url: z.string().url(),
      published: z.string(),
      abstract: z.string(),
    })),
  }),
});

// data is fully typed and validated
console.log(data.results[0].title);
docs.zapfetch.comv1.4
§Quickstart — 5 minutes2 min
§Authentication & API keys5 min
§Schemas & typed extraction4 min
§Streaming responses (SSE)3 min
§Batch requests & async jobs6 min
§Rate limits & error codes3 min
§SDKs — JS, Python, Go2 min
Full docsRead
All systems operational

Fetch the web. Now.

10,000 free requests a month, no credit card. Production SDKs ready today.

10k free / monthNo credit cardSOC 2 in progressEU & US regions