Context.dev cracked the top spot on Product Hunt this past July 2—a week, maybe less, since the company officially opened its doors. The YC-backed API racked up 955 upvotes and a 4.9 rating from 8 reviews, mostly from developers who'd grown weary of the entire web scraping song and dance: proxies that fail, JavaScript that doesn't render, pricing pages that AI agents somehow can't parse without throwing errors.
The pitch, when you strip it down, is simple. One API. Any website in, clean structured data out. Data that large language models can actually work with.
Yahia Bakour founded the company. He'd spent years at Amazon working on checkout infrastructure, then moved to Sunrun as a principal engineer. Somewhere along the way, he watched too many teams light engineering hours on fire just to build web scraping plumbing—time they could've spent on the product they were actually trying to ship. "We give AI agents realtime web context," the startup's YC profile reads, listing Context.dev in the Summer 2026 batch. By launch day, the company's website displayed logos for Klarna, Similarweb, Mintlify, and around 280 other outfits—though logo walls don't always equal confirmed paying customers.
The Mechanics of It
Context.dev sells a suite of endpoints designed to convert web pages into formats AI systems can stomach. The core scraping API spits out clean GitHub-flavored Markdown or raw HTML for one credit per request. It handles PDFs through OCR, Excel files, nested iframes. A separate crawl endpoint follows same-domain links as deep as 500 pages—handy if you're trying to ingest an entire documentation site or product catalog without writing custom logic.
Then there's the structured extraction API, which launched May 31. Pass it a JSON Schema, get back typed data that matches. Ten credits per successful call. Developers can feed in Zod or Pydantic schemas, add instructions, and receive structured objects instead of walls of text. Anti-bot challenges and JavaScript rendering happen automatically, with no surprise charges for "premium proxies" or stealth mode—the kind of upsells that have become standard at some competitors.
Other endpoints grab screenshots for five credits, pull brand assets like logos and color palettes for ten, retrieve company firmographics. A monitoring feature announced July 9 watches websites for changes as often as every ten minutes, pinging developers when pricing pages or product specs shift.
Pricing That Doesn't Move the Goalposts
The credit model starts with a free tier: 500 credits monthly at 30 requests per minute. Developer plans run $25 for 10,000 credits and 60 RPM. Pro costs $149, delivers 200,000 credits and 300 RPM. Scale hits $499 for a million credits at 700 RPM. Overages drop from $15 per 10,000 credits at the Developer tier down to $7 at Scale. Enterprise plans begin at two million credits with custom rate limits and service-level agreements.
The pricing page makes a point of the "no hidden multipliers" line—no extra fees for bot protection or JavaScript execution. A simple Markdown scrape costs one credit whether the site throws CAPTCHAs or not. It's a pitch aimed squarely at teams running high-volume retrieval-augmented generation pipelines who've been burned before by usage-based pricing that escalates without warning.
Early Traction, Familiar Faces
Daily.dev uses Context.dev to keep company metadata current, according to a case study published roughly five months before the July launch. Mintlify and DocsBot show up in testimonial quotes on the homepage, though logo walls don't always equal paying customers—a familiar ambiguity in early-stage marketing. The "Launch HN" post on Hacker News from July 10 pulled 116 points and 86 comments. Solid traction for a YC batch member selling an API to the agent-building set, though not a breakout viral moment.
The competitive field is crowded. Firecrawl handles RAG-scale Markdown ingestion. Diffbot builds entity-relationship graphs. Apify offers a marketplace of over 20,000 pre-built scrapers. ScraperAPI focuses on proxy management. Context.dev carves out space by emphasizing structured extraction over raw HTML access or pre-assembled knowledge graphs. A July 9 comparison post from the company itself—biased, naturally—lays out the tradeoffs: Firecrawl gives you open-source flexibility and browser interaction, Diffbot excels at relationships, Apify provides breadth. Context.dev bets on real-time structured data through a single API with integration hooks for agent frameworks.
Security Theater or Real Compliance?
Context.dev secured SOC 2 Type I certification on May 14. Type II was in progress as of mid-July, according to the company. The Trust Center lists Cloudflare, Firebase, Google Cloud Platform, and Hetzner as subprocessors. Rate limits scale with plan tiers. Caching controls via maxAgeMs parameters let teams avoid redundant calls. SDKs cover TypeScript, Python, Ruby, Go, and PHP, with documentation hosted—perhaps inevitably—on Mintlify.
The startup rebranded from Brand.dev to Context.dev back in March, broadening the scope from brand asset retrieval to general web data infrastructure. An older AI Website Query API from June of last year has been deprecated in favor of the newer /web/extract endpoint, which offers more granular control over schema and extraction logic.
Bakour's team grew from solo founder to at least three people around launch, with Nikhil Radosevich and Beer van der Drift coming aboard in early July. The YC company page listed team size at one as of mid-year—the kind of data lag common on startup directories that scrape their own information infrequently. Beyond YC's standard $500,000 deal—structured as $125,000 for 7% equity plus $375,000 on an uncapped MFN safe—no additional funding rounds have surfaced publicly.
What You're Actually Buying
The value proposition boils down to elimination. No more stitching together Puppeteer, proxy rotators, and parsing logic. Call an endpoint, get Markdown or JSON back. The credit model makes costs predictable. The SOC 2 cert addresses enterprise procurement hurdles that often kill API adoption at larger companies before it starts.
Whether Context.dev becomes foundational infrastructure or just another tool in the agent-building toolkit depends on extraction quality at scale. Some Reddit threads from earlier in the year mention trouble with self-hosted extraction configs and mixed results on complex sites, though those posts predate the May 31 launch of the current API. The Product Hunt momentum and Hacker News engagement suggest demand is real—developers building agents need web context, and they'd prefer to pay per call rather than maintain scrapers themselves.
For now, Context.dev delivers what's on the label: an API that turns websites into data. Execution and market response will write the rest of the story.
