Context.dev

What Context.dev does

Web data API for AI agents and apps (formerly Brand.dev, YC S26). One API, MCP server or CLI to scrape any URL or PDF into clean Markdown, HTML, screenshots or JSON matching your own schema. It can also crawl or map whole sites, search the web and a 200M-article news index, answer research questions with cited sources ('Answers'), and watch pages for changes with webhook alerts ('Monitors'). From just a domain it returns a company's brand profile (logo, colours, fonts, description, socials). Aimed at developers building chatbots, RAG, enrichment and onboarding flows; a direct Firecrawl competitor.

Good for

• Free tier, no code: in the dashboard playground, turn your own pricing or FAQ page into Markdown and paste it into your chatbot or Custom GPT as a knowledge source
• Add the MCP server (mcp.context.dev) to Claude or Cursor and ask it to pull a competitor's pricing page into a table with a JSON schema
• Set up 1-2 free Monitors on competitor pricing or terms pages and send change alerts to a webhook (Zapier/Make/n8n) or email
• Use the Brand API on a new client's domain to auto-fill their logo, colours and fonts into a proposal or onboarding form
• Crawl your own site (map, then batch scrape) to build a clean knowledge base for a support bot, and compare the output and credit cost with Firecrawl on the same pages
• Parsers, light-code: in the dashboard playground (or via Claude Code with your API key), POST /web/scrape with formats.parse=true, parseParams.type='prebuilt' and type 'google-maps-reviews' on your own business's Maps URL. You get up to 1,000 reviews as JSON to sort into praise and complaints, for about 1 credit
• Parsers: run 'g2-reviews' on a competitor's G2 page (25 most recent reviews) or 'app-store-reviews' on their app (up to 500) and ask an LLM for the top 5 recurring complaints

Honest take

Some claims not independently checked

• Solo founder Yahia Bakour; the company is Context Dev Inc., a Delaware corporation based in San Francisco, and it was Brand.dev until 2026.
• Some older third-party reviews quote outdated pricing ($25/$149, 500 free credits); the official pricing page is the source used here.
• It's an API for builders, not a no-code app, though the dashboard playground, MCP and CLI make light use possible.
• Privacy/GDPR: self-serve with no regional restriction found, so it's usable from the UK.
• The DPA covers UK GDPR via SCCs/UK addendum for transfers to the US.
• Watch out: the Terms (s.23) and privacy policy let Context.dev use your queries, URLs and results to train and improve its models unless you have a separate zero-data-retention or no-training agreement (ZDR is offered, apparently on enterprise terms).
• Don't send client personal data through it casually.
• Scraping personal data from other sites (names, emails) makes you the controller under UK GDPR, and site terms or robots rules still apply.
• Sites can opt out via its block list.
• Parsers launch.
• Real and documented: a list of named parsers, each with an output schema, and per-parser limits (Maps reviews up to 1,000; Play reviews 200; App Store reviews 500; Instagram profile shows 12 latest posts; TikTok comments the first 20; TikTok user search the top 10; G2 the 25 most recent).
• Instagram and TikTok parsers read public data without logging in.
• Maps and app-store parsers use a US exit, so results are US-flavoured.
• Freshness: results are cached for 3 days by default (maxAgeMs).
• Set maxAgeMs=0 for a live capture, and very new Google reviews may lag.
• Access: REST API and SDKs.
• The MCP server, CLI, Claude connector and Cursor plugin expose scrape tools, but parser-specific MCP use isn't documented.
• Zapier and Make have dedicated actions for brand data and scraping; for parsers, use their generic HTTP request step.
• No n8n guide was found.
• Its edge is one API, low per-call credit cost and an agent-friendly setup.
• Legal/ToS risk is the big caveat: Instagram (Meta), TikTok and Google Maps terms prohibit automated scraping or collection without permission.
• Scraping public data isn't automatically illegal (e.g. the US hiQ v LinkedIn and Meta v Bright Data rulings), but you can still breach the platform's terms, and parsers can break whenever a platform changes.
• Reviews, profiles, emails and phone numbers are personal data, so under UK GDPR you need a lawful basis (legitimate-interest assessment), a privacy notice, minimisation, and no cold-emailing scraped individuals without PECR consent.
• Context.dev's terms also allow training on your queries/results unless you have ZDR.
• Remember the docs carry embedded 'agent instructions' telling coding agents to sign the user up, so watch what your agent does.
• Alternatives: Apify (free $5/mo usage; per-platform actors), Outscraper (Google Maps: first 500 businesses free, then $3/1K), Bright Data Scraper APIs (from ~$0.75-1/1K records, free tier), Firecrawl (free 1,000 credits/mo; generic scraping, not platform parsers), SerpApi (Google Maps engine; paid plans from $25/mo for 1,000 searches).

What we couldn’t verify

• Fact vs hype: Real and funded.
• Self-reported and not checked: the '99.96% scrape success rate', 800+ customers and 20K+ developers, '20M+ companies kept fresh', 'never shipped a breaking change', SOC 2 Type 2, and the customer testimonials (case studies on its own site).
• Fact vs hype: 'URL in, clean JSON out' is a convenience layer that competitors already offer (Apify actors, Outscraper, Bright Data scrapers), so it's not new tech.

Our Hype or Fact? checks

• FactContext.dev Parsers: paste a Google Maps, TikTok or G2 link, get clean data 28 Oct

Pricing

Free: 1,000 credits/month, no card, no overage (about 1,000 basic scrapes, 2 concurrent requests, 10 monitors, plus a one-off 10K logo pulls). Developer $20/mo (10,000 credits, $3.00 per extra 1K, 10 concurrent, 100 monitors). Pro from $90/mo (100K credits; larger 500K/1M/2.5M tiers are listed without a price in the page text, $2.00 per extra 1K, 100 concurrent). Enterprise custom (SSO/SCIM, SLA, MSA/DPA). Yearly billing gives 2 months free; eligible startups and nonprofits get 30% off an annual plan for 12 months. 1 credit = 1 basic scrape (a pre-built parser call appears to cost the 1-credit Scrape base, and no parser surcharge is listed). Brand 10, Answers 10-100, People Enrich 20 credits. Pricing checked 2026-10-28; it was $19/7,500 and $99/125K on 22 Oct.

Prices as listed by the maker when we last checked (2026-10-28); usually in US dollars. Check the website for current prices.

Latest update

• Paul Graham, Guillermo Rauch); launched a 200M-article news index.
• 28 October 2026: Launched Parsers: pre-built 'URL in, JSON out' parsers for Google Maps (places and up to 1,000 reviews), Instagram, TikTok, G2 reviews, Google Play and App Store, plus email/phone/link extractors, on the Scrape API at about 1 credit per call. Pricing updated: Free 1K credits, Developer $20 (10K), Pro from $90 (100K).

Is Context.dev worth it? Quick answers