Ai-Q Labs

Extract tables from PDFs and rejoin the ones a page break cut in half. Reads ruled and borderless tables, drops the header the file repeats on every page, names each column type, and returns one clean grid per table as JSON and CSV.

Audits App Store listings for what the store never rechecks. Finds developer links that lead nowhere, domains behind them that lapsed or changed hands since the app was last updated, apps absent from every EU storefront while still selling elsewhere, and listings untouched for years.

Extracts article text and Markdown, and reports three things other extractors stay quiet about: a paywall that cut the body short, an article that continues on later pages, and tables and links that plain-text output destroys.

Check a list of domains in one run: DNS records, mail provider from MX, SPF/DMARC, HTTP status and full redirect chain, SSL certificate expiry, and WHOIS/RDAP registration - plus a plain-English issue list per domain. No API keys.

Chrome quietly stops distributing extensions your browsers keep running. Give this any list of extension IDs and it reports listings Google has pulled, extensions nobody has updated in years, and privacy policy, website and contact links that are dead - including domains anyone can now register.

Pulls every URL the Internet Archive captured for a site, checks each one against the live site today, and suggests where the dead ones should redirect. Catches the two losses other tools call healthy: 200s that say "not found", and 301s that dump every old path on the home page.

A bulk domain expiry checker and availability checker over RDAP. Free, registered or reserved; where the name sits in the delete lifecycle (grace, redemption, pendingDelete); when it could drop; whether it can be transferred today; and what every EPP status means.

Checks GitHub repositories against what the package registry claims. Catches what a registry never reports: a package served with no deprecation notice whose repository is archived, a repository that moved owner while requests keep returning 200, and a licence that changed after you installed it.

Google Play requires a support email on every listing and never rechecks it. This audits them: addresses whose domain bounces, domains nobody has registered at all - anyone can take that address - developer websites that lead nowhere, and apps left behind the target API level requirement.

Hacker News recommends links forever and never rechecks them. Point this at the front page, or at every story above 300 points from 2013, and it reports the dead links, the redirects that answer 200 OK while the article is gone, and the domains nobody has registered - with an archived copy of each.

Bulk URL status check that also grades the redirect: 301s that silently drop the path to the homepage, loops and long chains, HTTPS downgrades, meta-refresh and JavaScript redirects, soft 404s answering 200, and noindex on live pages. One row per URL with a plain-English issue list.

Checks whether a US nonprofit's tax-exempt status is still valid. Form 990 data alone cannot tell you: an organisation can carry a decade of filings and the deductible flag while the IRS revoked its exemption. Reconciles the 990 record against the IRS auto-revocation list and Publication 78.

Bulk PDF audit before OCR or text extraction: which pages are scanned and need OCR, which files are encrypted, what the metadata leaks (author, local paths), embedded JavaScript and attachments, PDF/A and PDF/UA conformance, unembedded fonts, and every link inside the document.

Convert PDFs to clean text and Markdown in the order a human reads them. Two-column pages are un-interleaved instead of read row by row, repeated running heads and page numbers are dropped, and words broken across line breaks are rejoined.

Government contracts already awarded, not open opportunities: who won, for how much, and how many bidders they beat, from the US, UK and EU registers in one schema. Flags contracts labelled openly competed that drew a single bid, ranks repeat winners per buyer, never fills a blank with zero.

Directories point listeners at a podcast feed forever and never recheck it. Reports episodes whose audio answers 404, feeds that died or moved, and shows whose lastBuildDate is a year fresher than their newest episode - 14.2% of 300 sampled shows, all looking active to anything reading that date.

Crawls a site from one starting URL and reports how visible it is to AI: which AI crawlers robots.txt admits, whether llms.txt exists, and what structured data and metadata each page offers. Give it a domain to crawl - for a fixed list of pages, use SEO Audit Tool instead.

Give it a list of URLs and it audits each page for on-page and technical SEO, plus the four checks general auditors skip: AI crawler rules, /llms.txt, redirects that throw the path away, and noindex via X-Robots-Tag. One row per URL - to crawl a whole site instead, use SEO & AI Visibility Auditor.

Shopify requires every app to publish a privacy policy and never rechecks the link. This audits the links a listing declares: policies that 404, policies that redirect to a homepage, domains nobody has registered - anyone can publish there - and dead developer and documentation links.

Detect what any website is built with - CMS, e-commerce platform, framework, web server, CDN, analytics, ad pixels, payments, support chat and consent tools - with the exact evidence for every match. Respects robots.txt. No API keys.

Decodes VINs and checks each vehicle against NHTSA's recall campaigns and its open defect investigations - the ones that have not produced a recall yet, and which a recall search therefore reports as nothing at all. Says plainly what is model-level and what is not.

Takes the dataset an audit Actor produced and opens a GitHub issue for each confirmed finding, through an MCP connector. It reads the tracker first, so running it again does not file the same finding twice, and it refuses to file anything the audit could not verify.

Find every sitemap a site publishes, expand sitemap indexes, and audit each URL: broken links, redirect chains, URLs blocked by the site's own robots.txt, XML errors, duplicates and invalid lastmod. One row per URL. Respects robots.txt.