Context.dev launches one API for agents to scrape and monitor the live web

Context.dev, founded by Yahia Bakour, launched on Y Combinator on Wednesday, August 5th, offering a single API that collects, structures, and continuously refreshes internet data for AI agents. The product, previously named Brand.dev, has over 400 customers including Mintlify, Passionfroot, and Rho, and competes with Firecrawl and Browserbase in the web data extraction market.

Yahia Bakour @mynameisyahia https://x.com/mynameisyahia officially launched Context.dev https://www.context.dev/ on Y Combinator on Wednesday, August 5th, pitching a single API that collects, structures, and continuously refreshes internet data for AI agents. https://x.com/mynameisyahia/status/2085065519515009397 https://x.com/mynameisyahia/status/2085065519515009397 The launch, detailed in a thread on X https://x.com/mynameisyahia/status/2085065519515009397 and a Y Combinator post https://www.ycombinator.com/launches/SPy-context-dev-live-web-data-api-for-ai-agents , is a distribution milestone for a product Bakour started building in 2025. Context.dev already operated before Wednesday and changed its name from Brand.dev on March 21st as Bakour expanded beyond logos and company metadata into broader web extraction infrastructure. Bakour entered YC's Summer 2026 batch as Context.dev's sole listed founder. He now says Context.dev has a team of four. Before starting Context.dev, Bakour led AI engineering work at residential solar provider Sunrun, worked on Amazon's checkout systems and built two acquired startups. His personal site https://yahiabakour.com/ identifies them as stock-alert app Stock Alarm and customer-feedback analysis product Essense.io. That history explains the product Bakour is selling. In the YC announcement, he said he repeatedly watched engineering teams assemble and maintain the same collection of browsers, proxies, crawlers, parsers, queues, retries and caches whenever an AI application needed current information from the web. Context.dev packages those components behind one API. Developers can send Context.dev a URL, domain, sitemap or file and receive Markdown, rendered HTML, screenshots, company information or JSON matching a specified schema. Context.dev also offers batch processing and monitors that watch pages, sitemaps or extracted fields and call a webhook when data changes. Its documentation https://docs.context.dev/ lists SDKs and integrations for scraping, structured extraction, document parsing, brand retrieval, screenshots and transaction enrichment. Context.dev is also trying to reduce integration work at the point where an agent first encounters the service. Bakour added a hosted Model Context Protocol server on August 3rd, allowing tools including Claude, Cursor, Codex, ChatGPT and VS Code to connect through OAuth. Context.dev also lets coding agents register for an account and obtain API credentials through an agent-readable setup file. Bakour says Context.dev is used by more than 400 customers, including Mintlify, Passionfroot and Rho. That figure is self-reported. Context.dev's customer pages https://www.context.dev/customers describe deployments ranging from turning GitHub repositories into branded documentation to monitoring source websites and enriching customer records with company details. Several of the usage figures attached to those examples also come from Context.dev or its customers. Context.dev says Murph uses the API behind a health database covering over 2 million foods, supplements and tests, while Executor uses it to discover documentation across more than 3,000 integrations. Dench attributes over $3 million in sales pipeline to a lead-monitoring workflow built with Context.dev. The market is already crowded. Firecrawl https://www.firecrawl.dev/ sells search, scraping, crawling and browser interaction for AI systems, while Browserbase https://www.browserbase.com/industry/ai provides managed browsers and agent infrastructure. Exa, Tavily and other web-search API providers compete for adjacent workloads involving retrieval, enrichment and fresh web content. Bakour is confronting that comparison directly. When an X user mentioned Context.dev alongside Firecrawl, Bakour challenged the user to load both API keys, run 10,000 production URLs in parallel and have an agent compare the results. The exchange included no benchmark data, leaving Context.dev's performance advantage unproven. Context.dev's immediate pitch rests instead on breadth, straightforward metering and the ability to consolidate web scraping with structured brand and company data. Context.dev's current pricing https://www.context.dev/pricing starts with 500 one-time free credits. Paid plans cost $25 a month for 10,000 credits, $149 for 200,000 and $499 for 1 million, with enterprise pricing above 2 million monthly credits. A standard page scrape consumes one credit, while structured extraction and several enrichment endpoints consume more. The YC launch gives Bakour a larger stage for a bet he began making months ago: AI developers will increasingly buy a maintained context layer instead of operating their own scraping stack. Context.dev must now prove that combining extraction, monitoring and company intelligence in one API produces better economics and reliability than assembling specialized providers.