Context.dev tools

The visual half of the brand record — logos, backdrops and the color palette — without the description, socials or firmographics

Resolve a listed company's ticker to its domain and brand record — logo and colors for a portfolio or watchlist UI

Turn a work email address into the employer's brand record — the signup-form enrichment call

Build a structured product/plan catalog from a company's site — name, price, currency, billing frequency, pricing model, features, audience and URL

Discover bounded lists of URLs on a domain without fetching page bodies.

Retrieve logos, colors, fonts, slogan, description, socials, address, headcount and industry for a domain

Resolve a company name to its domain and return the full brand record — logos, colors, description, socials, industry

Resolve an ISIN security identifier to the issuing company's domain and brand record

Find ranked web pages by query, with optional domain filters.

Return one page as Markdown, HTML, images or a screenshot.

Enumerate a site's URLs from its sitemaps so a crawler, index, or enrichment job starts with complete coverage. Sitemaps are discovered from the domain automatically, or a single sitemap URL can be crawled explicitly. An RE2 pattern filters which URLs are returned (and counted against maxLinks), so a section of a large site can be listed cheaply before any page is scraped.

Turn additive identity clues — social profile URLs, a name plus company, a work email, education, or location — into the best person candidate with an identity match score from 0 to 100. Returns name, current role with organization and dates, full experience and education history, skills, bio, avatar, location, and social and website URLs, or a clean not_found result to branch on. All supplied clues are considered together, so more clues mean a stronger match. Suited for CRM enrichment, lead qualification, signup intelligence, and candidate research. A not-found person is free; free and disposable email addresses are rejected upstream at no charge.

Turn a single product page into structured commerce data: name, price and currency, billing frequency and pricing model, category, features, target audience, tags, imagery, and SKU. Context.dev first classifies the URL, so a non-product page is reported as such rather than hallucinated into a product. Built for catalog ingestion, competitive pricing monitors, and merchandising workflows.

Point Context.dev at a website, hand it a JSON Schema plus optional instructions, and get back typed data ready to drop into a product, agent, or workflow — no per-site scraper. Context.dev prioritizes the internal links most likely to answer the schema, reads up to maxPages pages (hard cap 50), and fills the requested shape. factCheck: true restricts every value to facts stated on the page (unsupported fields come back null) while the default allows reasonable inference for derived fields such as ideal customer or competitor analysis. Typical uses: lead enrichment, pricing and packaging capture, hiring signals, compliance evidence.

Turn a file into LLM-ready GitHub Flavored Markdown: point file_url at any public https link (up to 25MB) and get back the parsed document as a downloadable Markdown artifact. Handles 60+ formats — PDF, DOCX, XLSX, PPTX, RTF, HTML, images, code files, and structured data (JSON, CSV, YAML, XML). Set ocr: true to extract text from images embedded in PDF pages (scanned contracts, faxes) — billed 1 credit per OCR-recovered page on top of the 1-credit base, and ocr_ran in the output tells you whether it ran. Markdown conversion mirrors the scrape family's switches (links, images, base64 shortening, main-content-only). The complement of the scrape endpoints: parse handles files you already have; scrape handles live web pages.

Warm the brand or styleguide cache for a domain before a lookup.

Give a research task and an example answer object; an agent picks web searches, page reads, and company or profile lookups, then returns json_content in your shape plus the source URLs it read and cited. 'fast' suits short factual tasks; 'ultra' (the default) researches deeper. Suited for company facts, market questions, and enrichment fields no single page holds.

Return the rendered DOM of a page when the application needs the HTML rather than cleaned text — custom parsing, archiving, QA, and extraction pipelines. Same rendering stack as the Markdown scrape (JavaScript execution, anti-bot bypass, premium proxies included), with options for main-content-only output, inline iframe rendering, animation settling, cache reuse, an extra post-load wait, and the proxy exit country.

Find the images on any page and return them with the element they came from, their type, and alt text — standard URLs, inline SVGs, data URIs, responsive srcset sources, CSS backgrounds, video posters, and embeds. Optional perceptual de-duplication keeps only the highest-resolution copy of each visually identical image. Useful for brand, catalog, and content pipelines.

Turn a single URL into clean Markdown optimized for prompts, RAG, search, and analysis. JavaScript rendering, anti-bot bypass, and premium residential proxies are handled server-side and included in the price. PDFs are fetched and parsed natively. Options control link and image preservation, base64 shortening, main-content-only extraction, iframe rendering, animation settling, cache reuse (maxAgeMs), an extra post-load wait (waitForMs), and the proxy exit country. The response carries the Markdown plus page metadata (title, language, canonical URL, author, site name, Open Graph, Twitter card, JSON-LD, and other meta tags).

Search indexed brands by name or domain for up to 10 lightweight matches.

Describe the companies you want in plain words and get a page of matches with name, domain, description, organization type, operating status, stock symbols, revenue range, industries, founding year, headquarters, headcount, and social profiles. Filter by headquarters country and employee range, and page on with next_cursor. 'fast' (the default) suits most lists; 'ultra' searches deeper and takes longer descriptions. Suited for account list building, market maps, and lead generation.

Find news coverage about one company, identified by name, website domain, stock ticker (optionally scoped to an exchange), or ISIN — or about a topic described in plain words (natural_language), when no single company is the subject. Use entity search for company monitoring; it verifies each article is about that company. Returns articles with stable ids, a story_id that groups syndicated copies of one announcement, url, headline, description, language, authors, lead image, publish date, article type (editorial, press_release, regulatory_filing, advisory), source publication, and a verified entity-relevance match with primary/secondary level and confidence. Supports publisher domain and country filters, language and article-type filters, a published-at date window, relevance or recency ordering, and cursor pagination. Suited for company monitoring, due diligence, deal sourcing, and news-driven enrichment flows.