Table of Contents
What Context.dev Actually Does
Context.dev is a web scraping and data extraction API that converts any URL into structured markdown, HTML, or JSON, with JavaScript rendering and full-site crawling via sitemaps. Its second product line, brand intelligence, extracts logos, colors, fonts, and company metadata directly from a domain — an unusual pairing most scraping APIs don't offer in one product.
Combining raw scraping with brand intelligence in one API
Most scraping tools stop at content extraction. Context.dev's brand intelligence layer — pulling logos, colors, and fonts from a domain — fills a genuinely separate need, useful for onboarding flows and AI agents that need visual brand context, not just text.
Getting Started: Account Setup
Go to context.dev and sign up with a work email to get 500 credits/month at 30 requests/minute — enough to prototype a real integration.
Basic Web Scraping
How to Scrape a URL
- Generate an API key from your dashboard
- Send a request to the scraping endpoint with your target URL
- Specify output format — markdown, HTML, or JSON — depending on what your downstream application needs
- For JavaScript-rendered pages, enable rendering in the request — this costs more but is necessary for content that doesn't appear in the raw HTML
Basic scraping costs 1 credit per call, making it cheap enough to prototype extensively on the free tier before committing to a paid plan.
Brand Intelligence Extraction
How to Extract Brand Data
- Send a request to the brand intelligence endpoint with a target domain
- Context.dev returns logo URLs, primary and secondary colors, font information, and company metadata
- Use this in onboarding flows — auto-populating a new user's brand kit from their company website, rather than requiring manual upload
Budget for brand extraction costing 10x a basic scrape
Brand intelligence calls consume significantly more credits than plain content scraping — factor this into your usage estimates before scaling a feature that relies heavily on brand extraction.
Structured JSON Extraction
How to Extract Structured Data
- Define a JSON schema describing the fields you want extracted (e.g., product name, price, description)
- Send the schema along with your target URL to the extraction endpoint
- Context.dev returns data matching your schema, ready to feed directly into a database or RAG pipeline without manual parsing
This is a strong fit specifically for grounding RAG systems and AI agents that need consistent, predictable data structure rather than raw unstructured text.
Understanding the Credit System
Context.dev bills by credits rather than a flat per-request fee, and different endpoints consume different amounts:
- Basic scraping (markdown/HTML/JSON): 1 credit per call
- Brand retrieval and structured extraction: 10 credits per call
Because costs scale differently by endpoint, forecasting spend gets harder as usage grows — track which endpoints your application calls most before assuming a flat cost per integration.
Full Workflow: Grounding an AI Agent
Here's a workflow for feeding live web data into an AI agent or RAG system:
Define your JSON schema for the data your agent needs
Structure the schema around what the downstream AI application actually consumes, not just what's available on the page.
Call the structured extraction endpoint against your target URLs
Test on a handful of representative pages before scaling to a full crawl.
Feed the returned structured data directly into your agent's context or RAG index
Consistent schema output means no manual parsing step between extraction and use.
Monitor credit consumption as you scale
Confirm actual usage matches your cost estimate before committing to a higher-volume paid tier.
Context.dev Pricing 2026
| Plan | Price | Includes |
|---|---|---|
| Free | $0 | 500 credits/mo, 30 req/min |
| Developer | $25/mo | 60 req/min |
| Pro | $149/mo | 300 req/min |
Start on the free tier — 500 credits is enough to prototype a real integration before committing to a paid plan.
Common Beginner Mistakes
- Not accounting for brand extraction's 10x credit cost. A feature relying heavily on brand intelligence burns through credits far faster than pure content scraping.
- Skipping JavaScript rendering for dynamic pages. Content that only appears after client-side rendering won't show up in a basic scrape without enabling this option.
- Building an unstructured scraping pipeline when a JSON schema would work better. If your downstream application needs specific fields, structured extraction saves a manual parsing step entirely.