About WebPeeks
Paste-in tools for reading the plumbing of a website, with the method behind every answer written down.
What this site is
WebPeeks is a set of free tools for checking the parts of a website that visitors never see: the HTTP response headers, the Content-Security-Policy, the TLS certificate, the DNS records that decide whether its email is trusted, its robots.txt, and the tags that control how it looks in search results and shared links. There is also a JWT decoder and a user-agent parser, because they answer the same kind of question about the requests going in and out.
How the tools work
Every tool is JavaScript that runs in your browser. There is no account, no server-side processing, and nothing you paste is transmitted or stored. That is also why the tools ask you to paste data rather than type a URL: a web page is not allowed to fetch another site's headers or certificate, and this site deliberately has no server to do it on your behalf. Each tool page gives the command that produces the right input.
Every page renders a complete worked example before any script runs. The example inputs use the
reserved example.com family of domains; the example certificate and keys were generated for
this site and sign nothing real.
Who runs it
WebPeeks is an independent site, run by one person who looks after a small fleet of static websites on a single Linux server: nginx configuration, TLS, DNS and email records for all of them. The checks here are the ones used on that fleet, written up so they can be run on anyone's site. The site is not affiliated with any hosting, security or email company. It is funded by advertising when ads are switched on (the privacy page says whether they are), and any link that pays a commission is labelled as one.
Methods and limits, tool by tool
Website report card
Runs the headers, CSP, email, robots.txt and meta tag checks below over what you paste and combines them: each area is scored out of 100 and the overall grade is a weighted average (security headers 30, CSP 20, email authentication 20, crawlability 15, meta tags 15) of the areas you gave input for. The area deductions and caps are listed on the page. The weighting is this site's own. Each finding carries a fix: a configuration line for nginx, Apache and Cloudflare, a DNS record, a robots.txt line or an HTML tag, written as a starting point to adapt. Limits: the limits of each underlying check, plus that it grades a single response and the homepage path only.
HTTP security headers checker
Parses raw Name: value lines, splitting a redirect chain into separate responses and
grading the last. Six headers are scored out of 100: Strict-Transport-Security 25,
Content-Security-Policy 25, clickjacking protection (X-Frame-Options or CSP frame-ancestors) 15,
X-Content-Type-Options 15, Referrer-Policy 10, Permissions-Policy 10. A is 90 or more, B 75, C 60, D 40.
This weighting is this site's own judgement, not a standard, and other graders will
differ. Limits: it sees one response, cannot tell whether that response arrived over HTTPS (HSTS only
counts over HTTPS), and knows nothing about the application behind the headers.
Content-Security-Policy evaluator
Parses the policy as the W3C Content Security Policy Level 3 specification describes: first
occurrence of a directive wins, keywords must be quoted, and fetch directives fall back to
default-src (with frame-src and worker-src following their longer
chains). The script checks follow the recommendations in Weichselbaum et al., CSP Is Dead, Long
Live CSP! (ACM CCS 2016): nonces or hashes with 'strict-dynamic' rather than host
allowlists. Limits: a static reading of the text. It cannot know what an allowed host serves, and
evaluates only the first of several comma-separated policies.
SSL/TLS certificate decoder
A DER (ITU-T X.690) reader written for this site walks the certificate structure in RFC 5280. Hostname matching follows RFC 6125 as browsers apply it (a wildcard covers exactly one leftmost label). Fingerprints are SHA-256 and SHA-1 of the DER bytes via the browser's Web Crypto API. Lifetime limits come from the CA/Browser Forum Baseline Requirements, including the reductions in ballot SC-081 (200 days from 15 March 2026). Limits: it does not verify the signature, build a chain to a trusted root or check revocation, and it judges expiry by your device's clock.
SPF, DMARC and DKIM record checker
SPF is checked against RFC 7208 (syntax, qualifiers, the 10-lookup limit in section 4.6.4, terms after
all, one record per domain). DMARC against RFC 7489, with the tags proposed in the IETF's
DMARCbis revision recognised. DKIM against RFC 6376 and RFC 8301; the RSA key length is measured by
decoding the public key. Limits: it cannot query DNS, so it cannot follow include or
redirect and its lookup count is a lower bound, and it checks records, not real signed
messages.
Meta tag and Open Graph checker
A tag scanner reads <title>, <meta> and <link>
from the pasted HTML, skipping comments and scripts. Open Graph requirements follow ogp.me; X card
types follow X's published card documentation. The title and description length guidance (about 60
and 160 characters) is a rule of thumb: Google truncates by pixel width, not characters. Previews are
drawn in this site's own style and are approximate; the share image is not loaded.
robots.txt tester
Implements RFC 9309: groups selected by product token (case-insensitive, merged when repeated, falling
back to *), * and $ in patterns, percent-encoding normalised,
longest match wins, and Allow wins ties. Limits: it does not apply crawler-specific behaviour beyond
the RFC, such as Google's fallback from Googlebot-Image to Googlebot, and
robots.txt itself is only a request to well-behaved crawlers.
JWT decoder
Decodes the base64url header and payload of a JWS (RFC 7515, RFC 7519), explains the registered claims, and converts NumericDate values to UTC dates. Algorithms are named from RFC 7518 and RFC 8037. It never verifies the signature, because that would need your key. Expiry is judged by your device's clock.
User-agent string parser
Pattern matching over the tokens browsers send, most specific first, with a list of well-known crawlers. It reports frozen values (macOS 10.15.7, Android 10 with model K, Windows NT 10.0 for both Windows 10 and 11) as frozen rather than guessing. Limits: any user agent can be faked, and browsers built on Chromium often cannot be told apart from Chrome.
What the tools are not
They are not a security audit, a penetration test or a guarantee of email deliverability. A clean result means the specific things listed on each page look right in the data you pasted. For anything that matters to your business, have the configuration reviewed by someone who can see the whole system.
Corrections
Standards change, and browsers change faster. If something here is wrong or out of date, the method is written on each page precisely so it can be checked; please say so and it will be fixed.