AI tool listing.
Context.dev has evolved well beyond its original Brand.dev roots. The current product is a managed web-context API for software and AI agents: one key provides scraping to HTML or LLM-ready Markdown, full-site crawling, screenshots, sitemap discovery, web search, schema-shaped extraction, brand/styleguide data, product and entity enrichment, company news and people enrichment.12 The core value proposition is not a new reasoning model; it is replacing the infrastructure between an agent and the live web.
For basic retrieval, the API can fetch rendered HTML or convert a page into Markdown and optionally strip page chrome so the model receives the main content rather than navigation/footer noise.23 JavaScript rendering, proxy escalation and anti-bot handling are managed by Context.dev instead of configured separately. The company says premium proxies, JavaScript rendering and anti-bot bypass are included in ordinary scraping credit pricing rather than billed as large multipliers.45
Crawling and extraction broaden the product beyond a page fetcher. The crawl endpoint follows same-domain links, supports depth/page controls, URL filtering and subdomain behavior, and returns structured success/failure metadata; current product pricing also exposes schema-based extraction that crawls a site and maps results into a caller-provided JSON schema.34 This is useful for RAG ingestion, research agents, competitive-intelligence pipelines and onboarding flows where the downstream system needs structured fields rather than arbitrary HTML.
The newer product surface adds recurring and asynchronous primitives that matter for real agents. Monitors can watch pages, sitemaps or extracted datasets on a schedule, keep run/change history and deliver signed HMAC-SHA256 webhooks when changes are detected.6 Exact-diff monitors capture literal changes; extract monitors semantically compare structured fields and can suppress cosmetic changes. The monitor feature is still described as beta, with semantic change detection being tuned, so it should not yet be treated as a mature guaranteed event source.6
Large jobs can move out of a synchronous agent turn. Context.dev’s current Cursor integration documents asynchronous batch workflows that can process a fixed set of up to 25,000 URLs or large crawls, expose progress/results and support cancellation/deletion.7 This is a practical architecture for an agent because a 10,000-page research job should not require one chat connection or one long-running request to stay alive.
Context.dev is also increasingly agent-native at the integration layer. Its hosted MCP surface exposes dozens of typed tools, and the company publishes integrations/plugins for Cursor, OpenClaw and Eve plus conventional SDKs and automation-platform integrations.781 The public product currently links TypeScript, Python, Ruby, Go and PHP SDKs, Make, Zapier and MCP.1 Public GitHub repositories make several SDK/client layers inspectable, though the core hosted scraping infrastructure remains a managed service rather than an open-source product.9
The data surface is broader than “web scraping.” Current endpoints include brand retrieval, colors, fonts and styleguides; product extraction; NAICS/SIC classification; transaction identification; news search; and people enrichment from an email, social profile URL or combinations of name/company/education/location.42 This can simplify an agent stack that would otherwise combine a scraper, proxy vendor, search provider, company-enrichment provider, logo API and monitoring service.
The breadth also creates an important architectural dependency. Context.dev becomes a central interpretation layer between the source web and the customer’s agent. Page cleanup, rendering, extraction, enrichment, caching and change detection are all product behaviors that can affect downstream answers. Teams that require reproducibility should therefore preserve source URLs, retrieval times and relevant raw/structured outputs rather than treating the API result as an immutable fact.
The company maintains a detailed changelog and is shipping quickly. During 2026 it expanded from brand intelligence into scraping, crawl/PDF/OCR, monitors, larger agent integrations and broader entity functions.3 Rapid development is a strength for capability coverage, but it makes version-aware integration and legal/documentation synchronization more important.
Context.dev now publishes unusually detailed pricing. Free costs $0 and provides a one-time allocation rather than a recurring monthly allowance: 250 API credits for personal-email signup or 500 for a work-email signup, plus 10,000 one-time Logo Link requests. Free has no overage and core API rate limits of 10 requests/minute for personal-email accounts or 30 for work-email accounts.4
Developer costs $25 per month and includes 10,000 credits/month, 60 API calls/minute, up to 50 concurrent monitors and two concurrent batches. Overage is $15 per additional 10,000-credit block.4 Pro costs $149 per month, includes 200,000 credits, 300 calls/minute, 500 concurrent monitors and five concurrent batches; overage is $9 per additional 10,000 credits.4
Scale costs $499 per month with 1,000,000 credits, 700 calls/minute, 10,000 concurrent monitors and 20 concurrent batches; overage falls to $7 per 10,000 credits.4 Enterprise starts beyond the published high-volume tiers and adds custom limits, volume pricing, SSO/SAML and SCIM, annual invoicing, MSA/DPA support, a dedicated support channel and a 99.9% uptime SLA.4
Annual billing is advertised as two months free compared with monthly billing, and startups/nonprofits can apply for discounts of up to 30% for one year.4 Context.dev does not currently offer ordinary one-time paid lookups as a standard checkout option, although one-off projects can be discussed with the company.1
The credit system is much more interpretable than many AI-credit schemes because the company publishes endpoint costs. Ordinary scraping, screenshots, page crawling and web search generally start at one credit per successful unit; structured extraction and several enrichment functions cost 10 credits; people enrichment costs 20 credits.4 Browser-action scraping can cost two credits and enriched image scraping five, so “one scrape = one credit” has documented exceptions even though ordinary retrieval remains simple.4
Failed or blocked requests are not billed, and paid plans can enable metered overage so production traffic continues after included credits are exhausted.4 Users can disable overage from the dashboard. This is a useful combination of predictable unit economics and continuity, although a poorly bounded crawler or agent can still create material spend because production overage is automatic when enabled.
The no-surcharge claim for stealth infrastructure is commercially meaningful. Context.dev says JavaScript rendering, proxy escalation, anti-bot bypass and premium proxies are included in ordinary scraping pricing rather than charged at a 5–25x multiplier.45 That simplifies estimation relative to vendors whose “hard site” requests consume many credits. It does not mean every site can be retrieved or that bypassing a technical block gives the customer legal permission to use the underlying content.
Monitoring has its own economic nuance. Creating/managing a monitor is free, but each actual run consumes the same credits as the underlying work: for example, one credit for a page/sitemap check or 10 credits for an extract monitor run.6 A monitor every ten minutes can therefore become a recurring API workload even though the monitor object itself costs nothing.
Brand data can also be cached. The pricing FAQ says Brand Retrieve can reuse cached data under a documented `maxAgeMs`, with a default around 90 days and a one-day minimum.4 That is efficient for logos/company metadata but means “live web” freshness varies by endpoint and cache controls. Buyers should distinguish page retrieval from enriched brand records when defining freshness requirements.
Overall, Context.dev’s pricing is a current strength. It exposes absolute plan prices, endpoint costs, overage blocks, rate limits, monitor/batch concurrency and enterprise thresholds instead of hiding commercial terms behind sales. The main cost risk is therefore workload design, not inability to understand the unit price.
The external evidence base is mixed but better than for many young infrastructure startups. There is no current G2/Capterra/TrustRadius-style corpus or independent exact-product benchmark. The sparse mature review coverage limits independent comparison with established scraping/data vendors.
Product Hunt provides the most useful current user evidence. The exact-product review page shows 4.9/5 from 15 reviews at capture time, with reviewers repeatedly praising clean Markdown, integration speed, SDK ergonomics, concurrency, responsive support and the ability to replace separate scraper/proxy/cleanup layers.10 Reported weaknesses include documentation/examples, brand-color extraction and uncertainty about how target-site robots.txt or Terms changes are surfaced to customers.10
The 4.9/5 Product Hunt rating should not be treated as a robust exact-product performance score. Thirteen of the 15 visible reviews are categorized as founder reviews, and the product has been launched/promoted on the same platform.10 The comments are valuable because several identify concrete production use cases and alternatives, but the sample is small and unusually founder-heavy; the rating is therefore best treated as qualitative user evidence.
Context.dev publishes significant first-party customer evidence. Its current customer page lists teams using the service for web search, crawling, branding, site monitoring and enrichment, including claims such as Bystreet completing more than one million scrapes without an issue and Adapt monitoring roughly 6,000 websites.11 These are useful workload descriptions and show a real production surface, but they are vendor-selected case studies rather than independent reliability statistics.
The public status page provides a more directly measurable operational signal. At capture, core scraping endpoints showed roughly 99.976%–100% uptime over the displayed 30-day window, with Markdown/HTML endpoints at 99.988% and several other endpoints at 100%.12 This is vendor-operated telemetry, not an audited SLA measurement, but it is more informative than a generic “high availability” claim and gives teams something concrete to monitor over time.
Security/compliance maturity has improved materially. Context.dev announced SOC 2 Type 1 compliance in May 2026 covering Security, Availability and Confidentiality, and its Trust Center shows SOC 2 Type 1 compliant with Type 2 still in progress.1314 Enterprise pricing separately advertises a 99.9% uptime SLA.4 Type 1 assesses design of controls at a point in time; it should not be represented as the longer operating-effectiveness evidence that a completed Type 2 report would provide.
The DPA is comparatively substantive. It defines Context Dev Inc. as processor for Customer Personal Data, requires processing on documented instructions, includes GDPR/CCPA commitments, sets subprocessor-change notice, contains security obligations and gives customers audit rights.15 Following cessation of relevant services, it states that Customer Personal Data should be returned or deleted promptly and in any event within 10 business days, subject to legal/backup exceptions.15
Context.dev also publishes a subprocessor page rather than forcing procurement teams to discover every vendor during sales. The list identifies processors, location/purpose and data categories, while the DPA provides a 14-day notification mechanism for intended additions/replacements.1615 That is stronger governance documentation than a generic privacy policy alone.
The general Privacy Policy was updated August 20, 2026. It describes account and usage data, third-party/analytics sharing, international transfers, data-subject rights and retention based on business/legal necessity rather than a universal fixed period.17 It says servers are in the United States and describes protections for international transfers. It also says the company does not currently respond to DNT browser signals but honors recognized opt-out preference signals such as GPC where applicable.17
The biggest trust weakness is inconsistency between the legal documents and the product now being sold. The Terms were also updated August 20, 2026, yet their opening description still frames the service mainly as a brand-data platform.18 More importantly, the user representations say a user will not access the Services through automated/non-human means, and the prohibited-activities section broadly bans automated system use, robots, scrapers and similar tools accessing the Services.18 Read literally across the defined “Services,” that conflicts with a product whose primary interfaces are APIs, MCP, SDKs and automation integrations for agents.
There may be an intended distinction between unauthorized automation against Context.dev’s own website and authorized API usage, but the current wording does not express that distinction cleanly.18 This is not a theoretical drafting nit: developers need contractual certainty that normal programmatic API/agent use is authorized. Context.dev should update the Terms so the automation restrictions clearly exclude documented API/MCP/SDK use.
The DPA shows a related scope lag. It says the Services are designed to retrieve brand/company information and do not require customers to submit end-user personal data, while the current pricing surface now includes a People Enrichment endpoint capable of enriching a person from email/social profile or name plus contextual attributes.154 The DPA still offers strong processor obligations, but privacy-sensitive buyers using people enrichment should confirm that the current contract/DPA schedule explicitly covers the real data categories and processing paths they intend to use.
Scraping compliance is intentionally allocated to the customer. Context.dev’s public materials state that customers are responsible for respecting target-site Terms of Service, robots.txt, rate limits and applicable privacy/data-protection law.25 Its engineering guidance itself recommends checking robots.txt, preferring official APIs where appropriate, avoiding authentication/paywall bypass without permission and minimizing collection.19 That is responsible guidance.
There is nevertheless tension between that guidance and the product’s technical sales message. Context.dev also advertises automatic proxy escalation, realistic browser fingerprints, JavaScript rendering, CAPTCHA/anti-bot capabilities and Cloudflare-style bypass as managed features.51 Technically bypassing a block and legally/contractually being entitled to collect the data are separate questions. Context.dev correctly places legal responsibility on customers, so adopters need their own allow/deny policy rather than assuming the API will enforce every target’s terms on their behalf.
The monitor feature introduces a similar compliance dependency. A Product Hunt reviewer specifically asked for clearer visibility into what happens if a target changes robots.txt or its Terms after a pipeline has already been built.10 Context.dev’s general materials say customers are responsible for those policies; the reviewed monitor documentation does not establish a universal automatic legal-policy change gate. Teams using long-lived monitors should therefore maintain their own target-policy governance for sensitive domains.
The evidence supports medium confidence. Context.dev has transparent pricing, a real customer surface, current status telemetry, SOC 2 Type 1, a public DPA/subprocessor regime and a modest exact-product user corpus. Confidence is held below high primarily by the absence of independent benchmarking/large review populations and by legal-document synchronization problems that are especially important for a scraping/people-data infrastructure vendor.
Context.dev is best suited to developers building AI agents, research systems, RAG pipelines and software features that need current web data but do not want to own browser automation, proxy rotation, anti-bot handling, page cleanup, crawl queues and enrichment logic.12 The value rises as the workflow needs more than one of those primitives.
It is particularly strong when a model needs compact readable content rather than a raw DOM. Product Hunt users specifically praise the Markdown cleanup for reducing navigation/footer noise before content enters an LLM context window.10 Research agents, coding agents that need current documentation and RAG ingestion pipelines are natural fits.
Teams that need recurring competitive or operational signals can use Monitors instead of rebuilding scheduling, baseline storage, diffing and webhook delivery.6 Extract monitors are especially useful when the desired signal is semantic—such as a price or plan changing—rather than any pixel/text edit. Because semantic monitoring is still beta, important workflows should retain validation and alert-quality measurement.
High-volume pipelines benefit from batch jobs and transparent scale tiers. A workload can move tens of thousands of URLs into asynchronous processing, while Pro/Scale plans expose significant rate limits and monitor concurrency.74 That is more suitable than trying to make a chat agent synchronously loop across a massive crawl.
Product teams that need company/brand context can reduce vendor sprawl by combining live page retrieval with logos, colors, fonts, styleguides, industries, products and other entity data.2 The same principle now extends to news and people enrichment, although personal-data use deserves a stricter compliance review.415
Context.dev is less attractive for teams that already operate a robust Playwright/browser/proxy/data-cleaning stack and need maximum provider-level control. Managed abstraction saves engineering effort but means rendering, cleanup, proxy selection, caching and enrichment behavior are delegated to another vendor.
It is also a weaker fit where legal policy must be enforced entirely by the vendor. Context.dev explicitly leaves target-site ToS, robots.txt, copyright and data-protection compliance with the customer.25 A regulated or litigation-sensitive organization should implement domain policies, provenance and review rules on top of the API rather than relying on technical retrievability as permission.
Organizations with sensitive personal-data workflows should review the DPA schedule against the current People Enrichment feature before deployment because the public DPA language still describes a narrower brand/company-focused data model.154
Finally, procurement/legal teams should resolve the Terms automation contradiction in writing for material deployments. A developer platform cannot reasonably be adopted at scale if documented API/MCP use appears to collide with generic prohibitions on bots/scripts/automated access.18
Products matched by shared tasks, audience, workflow, product type and capabilities.