Better Fetch
Reliable web retrieval for AI agents

When ordinary fetch stops, Better Fetch keeps going.

The reliable web retrieval layer for AI agents when ordinary fetch and generic scrapers stop working. Direct HTTP first, a real browser when needed, and metered residential recovery only as a last resort.

Start freeTry a retrieval

Connect with your Better Fetch account

In Claude, open Settings → Connectors, add a custom connector, paste this URL, then sign in. No API key to copy.

https://betterfetch.co/api/mcp

Claude Code power user

/plugin marketplace add better-fetch/claude-plugins
/plugin install better-fetch@better-fetch

Adds retrieval skills and a scraper subagent alongside the MCP connector.

OAuth sign-in50 free credits each monthFull setup guide

Ready-made capabilities

Ask for the data. Skip the scraper code.

Browse all 92 tools
Lead generation

Google Maps Search

Search Google Maps for places and get structured leads: name, address, rating, review count, category, phone, and website. The classic lead-list workflow, one call, no Maps API key.

Returns
nameratingreviewsphonewebsite
Examplecoffee roasters in Sydney
~1 credit/runView
Social

Instagram Scraper

Extract public Instagram profiles, posts, reels, comments, story highlights, highlight media, and speech transcripts, or raw public profile embed HTML.

Returns
urlauthor_idauthor_nameauthor_usernameauthor_verified
Exampleopenai
~2 credits/runView
Social

TikTok Scraper

Extract public TikTok profiles, follower/following lists, explicitly self-declared audience-country samples, videos, indexed search results, transcripts, comments, or replies.

Returns
avatar_urlbiodisplay_namefollower_countfollowing_count
Exampleopenai
~2 credits/runView
Social

Reddit Scraper

Inspect public subreddits, posts, search, comments, and Reddit-hosted video transcripts through official JSON and VTT surfaces.

Returns
titleurlauthorcreated_utcdomain
Examplesubreddit_details
~1 credit/runView
Social

Instagram Profile Scraper

Extract public Instagram profile metadata from a username or profile URL, including bio, links, avatar, verification, privacy, follower and following counts, media count, and recent visible posts.

Returns
titleurllink_type
Exampleopenai
~1 credit/runView
Search

Google Search Results

Structured Google search results — titles, URLs, and snippets — via the stealth browser engine. Optional country routing for localized SERPs. No SERP API subscription needed.

Returns
titleurlsnippet
Exampleweb scraping api
~1 credit/runView

Agent-native retrieval

One connector. The right fetch strategy.

Better Fetch gives the model both the tools and the operating rules to retrieve the web reliably, without blindly retrying expensive browser calls.

01
Escalates only when needed
Starts with fast HTTP, moves to Chromium for rendered pages, and can use residential egress when a target blocks direct traffic.
02
Keeps the browser state
Account-scoped sessions preserve cookies, localStorage, browser identity, and cache across an agent's multi-step work.
03
Returns data, not browser noise
Fetch pages, discover the APIs behind them, extract structured JSON, capture screenshots, or run a ready-made scraper.
04
Turns pages into documents
Normalize retrieved HTML into clean text, Markdown, metadata, canonical links, and images without a second network request.

MCP is the product. REST is the escape hatch.

Most people should connect Better Fetch directly to their AI. Developers can still use the same engine through the versioned API, with the same keys, sessions, metering, and fetch outcomes.

Connect your AIUse the REST APITry clean documents

Start with 50 free credits

No card required. Upgrade when your agents need more retrieval capacity.

Free
$0 / month
  • 50 credits / month
  • 1 stored browser session
  • OAuth or API key
Start free
Starter
$19 / month
  • 25,000 credits / month
  • 10 stored browser sessions
  • Screenshots
Start free first
Pro
Popular
$49 / month
  • 100,000 credits / month
  • Residential recovery
  • 50 stored browser sessions
Start free first
Scale
$199 / month
  • 500,000 credits / month
  • Everything in Pro
  • 250 stored browser sessions
Start free first

Ordinary retrieval costs 1 credit. Pro and Scale unlock residential recovery; each residential attempt that actually occurs costs 25 credits. Cache hits and coalesced followers stay at 1 credit.