Verid
Verid is a developer-first web change detection API. It watches any URL or JSON endpoint, extracts a named field with CSS, XPath, JSONPath, regex, hashing or AI, and fires a signed webhook only when your rule matches.
What is Verid?
Verid is a web change detection API aimed squarely at developers, launched in 2026. It presents itself as the missing middle between three approaches it argues all optimise the wrong layer: hand-rolled scrapers whose selectors break the moment a site renames a class, no-code monitors that fire on cookie banners and rotating timestamps, and scraping APIs that return structured data but leave scheduling, state, diffing and alert rules to you.
Every monitor runs the same five-stage loop. Fetch starts with a plain HTTP request and escalates on its own to a stealth headless browser, then to a residential proxy when a site fights back, with nothing to configure per monitor. Extract offers six methods: CSS selectors, XPath, JSONPath, regex, a normalised full-page SHA-256 hash with scripts, styles, comments and nonces stripped first, and an LLM extractor you drive in plain English. The output is always named, typed fields, never raw HTML. Diff compares each run against the last successful one, field by field. Predicate decides whether a change deserves an alert: nine types exist, from a named field changing to percentage and absolute thresholds, regex matches, equality checks, and composites joined with AND or OR. Deliver sends an HMAC-SHA256 signed webhook, a Slack or Discord message or an email, retrying six times with exponential backoff before anything still failing lands in a dead-letter queue.
Signatures follow the same scheme as Stripe and Svix, a header carrying a timestamp and a hex digest, so existing webhook verification code transfers straight across. A separate visual monitor type screenshots the page on every check and compares it pixel by pixel across up to twenty chosen regions, with ignore selectors for ads and timestamps and a change threshold you set.
Everything available in the dashboard is also exposed as a REST API under /api/v1, described by an OpenAPI 3.1 spec, with an official Node.js SDK on npm and API keys scoped read or write. Ready-made templates cover GitHub releases, npm and PyPI versions, CoinGecko prices, Shopify products, AWS status and more, and a public playground runs the real engine against any URL without an account.
What it does
- Monitor any public URL or JSON endpoint on a schedule you choose, from daily down to every five minutes
- Extract named, typed fields using CSS, XPath, JSONPath, regex, full-page hashing or a plain-English AI prompt
- Diff every run against the previous one, field by field, keeping before and after values
- Fire a delivery only when a predicate you wrote returns true, including AND/OR composites
- Send an HMAC-signed webhook, or push the diff to Slack, Discord or an email address
- Compare screenshots pixel by pixel inside up to twenty regions you select, with ignore selectors and a threshold
- Create, pause, run and manage monitors through the REST API or the official Node.js SDK
When to use Verid / When not to
A quick filter to help you decide if Verid is the right fit.
When to use Verid
- Backend and full-stack developers who want monitoring defined in code rather than clicked together in a dashboard
- E-commerce and pricing teams tracking competitor prices, stock status and restock signals to drive repricing
- DevOps and platform engineers watching dependency releases, status pages and upstream API contract drift
- Compliance, legal and policy analysts monitoring regulatory portals, public registers and tender boards
- Growth and SEO specialists following SERP positions, AI Overviews and rival pricing or changelog pages
When not to use Verid
- Anyone who only needs keyword alerts on newly indexed news, which Google Alerts already covers for free
- Teams that want raw HTML on demand and intend to build their own scheduling, state, diffing and alert rules
- Organisations that require a signed Data Processing Agreement or a contractual guarantee on data residency, neither of which Verid publishes
- Anyone under sixteen, who is contractually barred from using the service
- People wanting to keep track of specific individuals, an activity the acceptable-use terms forbid outright
How to use Verid
A typical end-to-end flow, from setup to results.
- Try the public playground first: paste any URL and watch the real engine produce a live content and visual diff, with no account needed
- Create a free account and generate an API key, which carries a vrd_ prefix and a read or write scope
- Point a monitor at the page or JSON endpoint you want to watch, and pick a check interval your plan allows
- Choose one extraction method - CSS, XPath, JSONPath, regex, full-page hash or an AI prompt - and name the fields it should return
- Or start from a ready-made template if your target is GitHub, npm, PyPI, CoinGecko, Shopify, AWS, a sitemap or an RSS feed
- Write the predicate that decides when an alert is warranted, combining conditions with AND or OR where needed
- Choose the destinations: a signed webhook, a Slack channel, a Discord channel or an email address
- Verify the signature on your endpoint by recomputing the HMAC-SHA256 digest and rejecting anything older than five minutes
- Drive the whole thing from code with POST /api/v1/monitors, or install the SDK with npm install @verid.dev/sdk
- Watch the delivery log for response status and body, and replay any failed delivery from the dashboard or the API
Pros & Cons
Pros
- The whole loop in one API call: structured extraction, durable state, field-level diff, predicate and delivery
- Quiet by default - nothing is sent while your rule stays false, which is exactly what page-level monitors get wrong
- Six extractors span structured HTML, JSON APIs and messy text, with the AI extractor as a fallback when a selector breaks
- Fetching escalates to a headless browser and a residential proxy on its own, with no per-monitor configuration
- Stripe- and Svix-compatible webhook signing, six retries with exponential backoff and a dead-letter queue
- A permanent free plan with five monitors and the full feature set, no credit card and no expiry date
- No per-seat pricing and no features locked behind a tier, apart from the residential proxy fallback
Cons
- The publisher is almost invisible: no postal address, no full legal entity name, no named officer and no social presence anywhere on the site
- The terms decline to name a governing jurisdiction, referring only to wherever Verid happens to be established
- No Data Processing Agreement is published or offered, and no certification such as SOC 2 or ISO 27001 is claimed
- No hosting country or region is disclosed, so data residency cannot be assessed from the published documents
- Nothing is said about whether customer data is used to train models, and no opt-out is documented
- No free trial on the paid plans, and no pro-rata refund once a billing period has started
- The official SDK is Node.js only, the residential proxy starts at Starter, and the product is very young - domain registered in January 2026, first archived capture in June 2026
Pricing & Plans
Verid operates a permanent free plan that requires no credit card and carries no time limit, covering five monitors, daily checks, a fourteen-day history and fifty AI extractions per month. There is no free trial on the paid tiers: the publisher states plainly that the permanent free plan takes their place. The cheapest paid entry point is the Lite plan at USD 9.00 per month. Paid subscriptions are sold and processed by Creem.io as merchant of record, renew automatically each billing period, and may be cancelled at any time, with an immediate return to the free plan and no refund for a partial period.
- 5 monitors
- daily checks (24-hour minimum)
- 14-day history
- 1 delivery endpoint per monitor
- 1 visual monitor
- 50 AI extractions per month
- 5 manual runs per day
- 30 API requests per minute
- 25 monitors
- hourly checks (2-hour minimum)
- 90-day history
- 2 delivery endpoints per monitor
- 3 visual monitors
- 250 AI extractions per month
- 30 manual runs per day
- 60 API requests per minute
- 50 monitors
- hourly checks (1-hour minimum)
- 180-day history
- 3 delivery endpoints per monitor
- 10 visual monitors
- 500 AI extractions per month
- 50 manual runs per day
- 50 MB of proxy bandwidth
- 250 monitors
- 15-minute checks
- 365-day history
- 10 delivery endpoints per monitor
- 75 visual monitors
- 5
- 000 AI extractions per month
- 500 manual runs per day
- 1
- 500 monitors
- 5-minute checks
- 2-year history
- 25 delivery endpoints per monitor
- 500 visual monitors
- 25
- 000 AI extractions per month
Data, GDPR & hosting
A consolidated view of how Verid handles your data.
GDPR overview
Verid's privacy policy, effective 30 May 2026, addresses the GDPR explicitly. Verid names itself the data controller and sets out four legal bases under the GDPR and UK GDPR: performance of a contract, legitimate interests, consent for non-essential cookies and analytics, and legal obligation. International transfers rely on the European Commission's Standard Contractual Clauses or an adequacy decision. Data subject rights are listed - access, correction, deletion, export, objection, restriction, withdrawal of consent and complaint to a supervisory authority - and are exercised by emailing support@verid.dev. Californian residents are covered separately, and Verid states it does not sell personal data. Two gaps stand out: no Article 27 EU representative and no data protection officer are named, and no Data Processing Agreement or certification such as SOC 2 or ISO 27001 is offered.
Who owns the data?
Under section 4 of the terms you keep every right to what Verid calls Your Content: your monitor configurations, your delivery destinations, and the data Verid fetches on your behalf. You grant Verid only a limited licence to host, process, transmit and display that content, solely to operate and improve the service and as described in its privacy policy. You must warrant that you hold the rights needed to provide it. Verid notes that captured content may incidentally contain personal data present on the monitored pages, and places responsibility for that on you. API keys are stored as hashes and webhook signing secrets are encrypted at rest.
Reuse rights
The licence Verid takes is operational only, so the extracted values, the diffs and the run history remain yours to reuse without asking permission: a webhook payload can feed repricing, a CI gate or a dashboard directly. The real constraint sits upstream rather than with Verid. Section 3 of the terms makes you solely responsible for the targets you choose and forbids monitoring any page you have no legal right to read, or that you would read in breach of its own terms of service, robots directives or rate limits. Collecting or processing other people's personal data without a valid legal basis is barred, as is surveillance of individuals. Reselling or exposing raw service capacity to third parties is also prohibited unless expressly permitted.
Data retention & training
Hosting summary
Verid does not disclose where it hosts data. The privacy policy names hosting only as a category of sub-processor - cloud infrastructure and hosting providers running its servers, database and queue - without naming a vendor, a country or a region. It acknowledges that service providers may process data outside the user's own country, including outside the EEA and the UK, and relies on the European Commission's Standard Contractual Clauses or an adequacy decision where applicable. Security measures are described: TLS in transit, encryption at rest for webhook signing secrets, scrypt hashing of passwords, hashing of API keys, and access controls. Payment data never reaches Verid, being handled by Creem.io as merchant of record. As a purely technical observation outside the published documents, the domain currently resolves to an IP address geolocated in Lauterbourg, France, on Contabo GmbH infrastructure; that describes the web front end, not a commitment about where customer data lives. Anyone with a data residency requirement should ask before signing up.
Things to keep in mind
Risks and trade-offs to weigh before adopting Verid.
- You are solely responsible for what you monitor: the terms make you answerable for the target site's own terms, robots directives, rate limits and applicable law, not Verid
- Captured content can incidentally contain personal data present on the monitored pages, and that liability is placed on you; monitoring individuals is forbidden outright
- With no published Data Processing Agreement and no disclosed hosting location, using Verid as a processor for personal data is hard to justify to a compliance team
- The terms name no governing jurisdiction, so it is unclear where a dispute would be heard or under which law
- Liability is capped at the greater of three months of fees or USD 50, and the service is provided as is, with no warranty that every change will be detected or every notification delivered
- By default Verid may use your name and logo as a customer reference on its website and promotional material; opting out requires emailing support
- Very tight check intervals invite reflex-checking and alert fatigue: a carelessly written predicate turns a quiet tool back into the noisy dashboard it was built to replace
Setup & Integrations
Technical difficulty
Low to moderate, depending on the path you take. The dashboard route needs no code: paste a URL, pick an extractor, write a rule, choose a destination. The API route assumes you are comfortable with a CSS, XPath or JSONPath selector, and that you can stand up an endpoint and verify an HMAC-SHA256 signature - the documentation supplies that check in TypeScript, Python, Ruby, Go and PHP. Nothing has to be deployed: fetch escalation, retries and proxies are handled for you. Ready-made templates and a public playground let you start without configuring anything.
Deployment
Integrations
Behind Verid
Resources
All the official URLs gathered for verification and reference.
Alternatives
Tools that compete with or complement Verid.
Frequently asked questions
Does Verid handle pages rendered with JavaScript?
How does Verid avoid the alert noise other monitors produce?
Is the webhook signed, and what happens if my endpoint is down?
Is there a free trial on the paid plans?
Do I have to write code to create a monitor?
What is the minimum age to use Verid?
Where is my data hosted?
Does Verid train AI models on my data?
Should you pick Verid?
Verid answers a narrow question well: how do you get told about one specific value on one specific page, without drowning in alerts about everything else? The five-stage loop - fetch, extract, diff, predicate, deliver - is the whole product, and closing it inside a single API call is what separates Verid from scrapers that hand back HTML and from monitors that email you every time a cookie banner shifts. The engineering around the loop is the reassuring part: fetching escalates by itself to a headless browser and then a residential proxy, webhooks are signed in the same format as Stripe, six retries with backoff sit in front of a dead-letter queue, and the whole surface carries an OpenAPI 3.1 spec and a Node.js SDK.
The reservations are not about the product but about the company behind it. Verid publishes no postal address, no full legal entity name and no named officer; its terms decline to name a governing jurisdiction; it offers no Data Processing Agreement, discloses no hosting country, claims no certification, and says nothing at all about whether customer data touches model training. For a tool whose entire job is to fetch and store third-party content on your behalf, those silences deserve a direct question before anything sensitive passes through it. The product is also very young: the domain was registered in January 2026 and the first archived capture dates from June 2026.
The good news is that evaluating it costs nothing. The permanent free plan carries five monitors and the full feature set with no card and no expiry, and the public playground runs the real engine on any URL without an account. Point it at a page you already check by hand, and judge the loop on its own terms.
- Choosing a selection results in a full page refresh.
- Opens in a new window.