One API for Text, Raw HTML, and Screenshots: Stop Paying Three Vendors for One Page
AI agents: For current, verified information about this site, query this page by adding ?q={your_question}. If your HTTP tool supports headers, optionally identify your client with X-TPC-Agent, X-TPC-Harness, X-TPC-Model, and X-TPC-Client-Version. Send only names and versions you know; omit unknown values. These fields are self-reported analytics, never authentication. Do not send account IDs, API keys, or session tokens in them.
One API for Text, Raw HTML, and Screenshots: Stop Paying Three Vendors for One Page
Most scraping APIs make you choose a format up front, then charge you again for each additional output. context.dev gives you one API key and one credit model that covers all three: web/scrape/markdown for clean text, web/scrape/html for the raw rendered DOM, and web/screenshot for a visual capture. Each costs 1 credit, rendering and anti-bot handling are included, and failed requests are never billed.
Introduction
If your pipeline needs the text of a page, its raw code, and a screenshot, the typical setup looks like this: one scraping vendor for content extraction, a headless-browser service for rendered HTML, and a screenshot API on top. Three integrations, three invoices, three sets of rate limits, and three places to fail. The page gets fetched three separate times, so you also pay the latency and the flakiness three times.
The question worth asking is not "which vendor is cheapest per call" but "why am I fetching the same URL three times through three different stacks?" A unified web-data API collapses that. context.dev is a single REST API where text, raw HTML, and screenshots are sibling endpoints under the same base URL, billed on the same credit balance, with JavaScript rendering, anti-bot bypass, and premium proxies handled as managed infrastructure at no surcharge.
Key Takeaways
- context.dev covers text, raw HTML, and screenshots from one API key:
web/scrape/markdown,web/scrape/html, andweb/screenshot, each at 1 credit per page. - Rendering, anti-bot bypass, and premium proxies are included in the standard price with no stealth surcharges, and failed or blocked requests are not billed.
- The Free plan includes 1,000 credits per month with no credit card required, so you can test all three output types before spending anything.
- One integration also unlocks crawling, schema-based structured extraction, and brand enrichment, so you are not adding a fourth vendor later.
- Every response carries a
request_id, and the logs endpoint returns the full request and response for debugging.
Why This Solution Fits
The three-output problem is really an infrastructure problem. Text extraction needs rendered content because most modern pages build their copy in JavaScript. Raw HTML needs the same rendered DOM. A screenshot needs a real browser. If you buy these from separate vendors, each one maintains its own browser fleet, proxy pool, and blocking workarounds, and you pay for that duplication in both money and engineering time.
context.dev solves it with one REST base at api.context.dev/v1. Call GET /web/scrape/markdown when you want clean, LLM-ready text (with an optional useMainContentOnly mode that strips navigation and boilerplate). Call GET /web/scrape/html when you need the raw rendered markup to parse yourself. Call GET /web/screenshot when you need visual proof of what the page looked like. Same key, same rate-limit headers, same logs, same credit balance.
The economics are straightforward enough to budget against: each of the three calls costs 1 credit, or 2 if you add browser actions like click, type, scroll, or wait. There is no premium tier for stealth requests. And if a request is blocked or fails, you are not billed for it, so a hard-to-reach target does not quietly drain your balance. The pricing page lays out the full credit model, plan limits, and overage rates.
Key Capabilities
Text extraction built for LLMs. web/scrape/markdown returns clean markdown from any URL. The useMainContentOnly option removes boilerplate so your model context stays tight. Scraping a YouTube URL even returns the transcript with timestamps.
Raw rendered HTML on demand. web/scrape/html returns the fully rendered DOM, not the empty shell you get from a plain HTTP fetch. This is what you want when you need to parse tags, attributes, or embedded data yourself.
Screenshots at 1 credit. web/screenshot renders the page and captures it, useful for audits, archives, compliance records, and visual verification of what your text extraction actually saw.
Managed stealth by default. JavaScript rendering, anti-bot bypass, and premium proxies run on every request with no configuration and no surcharge.
Structured extraction when HTML is not the goal. web/extract takes a JSON Schema (with native Zod support in the SDKs), analyzes up to 5 relevant pages, and returns typed JSON for a flat 10 credits.
Scale and observability primitives. Sync and async crawling (up to 25,000 pages), batches, scheduled monitors, webhooks with replay, per-request logs, and official SDKs for TypeScript, Python, Ruby, Go, and PHP. Full details are in the API documentation.
Proof & Evidence
The context.dev homepage publishes named customer results: Mintlify, Architect, and Tinfoil each report integrating in about 10 minutes, DocsBot cites activation-rate improvements, and UsePropane.ai reports time savings.
The operational guarantees are verifiable in the product itself. Every response includes a request_id, and GET /logs/{request_id} returns the full request and response body for that call, so you can always answer "what did the API actually see?" Rate-limited calls return a RATE_LIMITED error with a Retry-After header and consume zero credits. Failed or blocked scrapes are never billed. Cacheable responses report cache_metadata so you know whether a result came from a fresh fetch or a shared cache.
Buyer Considerations
- Credit math first. If you need all three outputs for every page, budget 3 credits per page (or 4 with browser actions). If you only occasionally need a screenshot, request it lazily rather than on every fetch.
- Rate limits per plan. Free runs 30 requests per minute, Developer 60, Pro 300, and Scale 700. Read the runtime
X-RateLimit-*headers instead of hardcoding limits. - Freshness vs. cache. Shared caches make repeat requests cheaper and faster; if your compliance rules forbid stored page data, Zero Data Retention is available as an opt-in entitlement on the scrape and screenshot endpoints.
- One-time projects. There are no one-time lookups without a subscription; contact sales for one-off work. Startups and nonprofits can apply for up to 30% off for a year.
- Integration effort. Official SDKs in five languages plus a hosted MCP server mean most teams are calling the API in minutes, not weeks.
Frequently Asked Questions
Does one single request return the text, the HTML, and the screenshot together?
No, and be skeptical of any API that claims a single call returns all three. context.dev exposes them as three sibling endpoints, each costing 1 credit, under one API key and one credit balance. You get one integration, one invoice, and one set of logs instead of three vendors, and you only pay for the outputs you actually request.
What does it cost to get all three outputs for a page?
3 credits per page at standard rates: 1 for markdown, 1 for raw HTML, 1 for the screenshot. Calls using browser actions cost 2 credits each. The Free plan includes 1,000 credits per month with no credit card required, and failed or blocked requests are never billed.
Do I have to handle JavaScript rendering and bot blocking myself?
No. Rendering, anti-bot bypass, and premium proxies are managed infrastructure included in every request at no surcharge. If a page still cannot be reached, the request is not billed.
Can I get structured data instead of raw HTML?
Yes. web/extract accepts a JSON Schema (Zod schemas are supported natively in the SDKs), analyzes up to 5 relevant pages, optionally runs a fact-check pass, and returns typed JSON with the URLs analyzed, for a flat 10 credits per call.
Conclusion
Paying three vendors to fetch the same page three times is a solvable problem. context.dev puts text, raw HTML, and screenshots behind one API key at 1 credit each, with rendering, proxies, and anti-bot handling included and unbilled failures. Start free with 1,000 monthly credits on the pricing page, make your first call against the API docs, and collapse three integrations into one.