The Internet's Context {API}

One API. Every piece of web context your agent needs.

One API Infinite Use Cases

Retrieve, extract, classify, and enrich web data in seconds.

Extract web content

Scrape any URL, or crawl an entire site, for clean markdown, rendered HTML, or extracted images.

Extract site map

Crawl and discover all page URLs from any domain via sitemap.

Retrieve brand data

Get logos, colors, and metadata by domain, email, name, or stock ticker.

Extract styleguide

Get the complete visual identity: colors, elements, and fonts from any website.

AI Query

Use AI to extract custom data from any website with structured output.

Get products

Extract product listings or individual product details with AI.

Capture screenshots

Generate high-quality viewport or full-page screenshots of any website.

Identify transaction

Match transaction strings to brands with optional location and MCC hints.

Extract web content

import ContextDev from 'context.dev';

const client = new ContextDev({ apiKey: process.env['CONTEXT_DEV_API_KEY'] });

// Get clean markdown from any page
const { markdown } = await client.web.webScrapeMd({
  url: 'https://openai.com/pricing'
});

// Get fully rendered HTML (JS executed)
const { html } = await client.web.webScrapeHTML({
  url: 'https://react.dev/learn'
});

// Extract all images from a page
const { images } = await client.web.webScrapeImages({
  url: 'https://dribbble.com/shots/popular'
});

// Crawl an entire site → markdown for every page
const { results, metadata } = await client.web.webCrawlMd({
  url: 'https://docs.stripe.com',
  maxPages: 50,
  maxDepth: 3,
  useMainContentOnly: true
});

What your agent can do with context.dev.

You have an agent. Here's what it can do now.

Read the web, live

Give an agent real-time access to any page. Scrape to markdown, render JS, extract images, so the agent reasons over the current web, not a frozen snapshot.

Ground RAG in fresh content

Keep your retrieval index current. Crawl a sitemap, convert every page to clean markdown, and pipe it straight into embeddings, with no brittle parsers in between.

Run deep research on demand

Answer "tell me about this company" with actual sources. Resolve a domain to a typed profile, then expand with custom AI queries pulled from the live site.

Personalize onboarding in a call

Make a new user feel seen from sign-up. One domain → company name, logo, colors, industry, auto-filled before they tab away.

Enrich any entity your agent sees

Turn the references in your prompts into structured data. One call hydrates a company reference with logos, colors, industry, socials, and firmographics.

Make sense of transaction noise

Resolve "SQ *BLUE BOTTLE 8xx" into Blue Bottle Coffee. Identify the brand, attach visuals and metadata, hand the agent something it can actually act on.

The Problem

Your agent is only as good as what it knows.

Most agent failures aren’t reasoning failures. They’re context failures. The agent didn’t have the right data, in the right shape, at the right time.

  1. Training data is frozen. Your model knows the web as of some cutoff. The web your user is asking about is newer than that, every time.
  2. Generic scrapers return noise. Raw HTML with nav, ads, and boilerplate forces your agent to guess. Tokens get burned on chrome, not signal.
  3. Entities need structure, not prose. An agent asked about a company shouldn't read a blog post about it. It needs logos, colors, industry, socials, already typed.

The context your agent can't get anywhere else.

Markdown from the live web, typed brand data on any company, and a one-line logo embed, all from the same API.

Web Extraction

Read the live web as markdown.

Brand Intelligence

Any domain, typed.

One <img> tag. Any logo.

No API call. No SDK. No backend. Just an image URL, with logos served from a global CDN in ~20ms.

FAQs

Is there a free tier?

Yes, Context.dev offers a free tier with 500 API credits and 10K Logo Link requests to test out both services.

Do you offer discounts for startups or nonprofits?

Yes! We offer a startup discount of up to 30% off for one year if you're still early-stage. Apply at https://www.context.dev/startup-discount.

How fresh is the brand data? How often is it updated?

Cached brand data is refreshed quarterly by default. Any brand older than 3 months is completely re-fetched when requested via the API.

What SDKs are available?

We provide official SDKs for TypeScript, Python, and Ruby.

Ship an agent that actually knows things.

Free tier, 10-minute integration, and the same API powering agents at Mintlify, daily.dev, and Propane. No credit card to start.