Choose a tool.
Start with your site
The llms.txt generator and crawler checker fetch public site information through our service. The FAQ generator creates markup from the questions and answers you enter in your browser.
llms.txt Generator
Enter a URL, get a valid, ready-to-publish llms.txt built from your site's real pages. Copy it, or download the file and upload it to your site root.
Open the generator → CHECKERAI Crawler Access Checker
Review robots.txt rules and compare homepage responses to requests using published crawler User-Agent strings.
Check your site → GENERATORFAQ Schema Generator
Turn your questions and answers into valid FAQPage JSON-LD that AI systems can read. Copy, paste, publish.
Build your schema →The standing promise: free to use, no account, no email address, and we keep nothing. We read the page you ask about, return the result, and that is the end of it.
How We Build,
Written Down
Not interactive — these are the plain-English explainers behind the stack and the standards we hold every build to. Read them if you want to know what you are actually buying.
WordPress
Our WordPress work since 2009: content models, online stores, integrations, and maintenance across platform changes.
WooCommerce
Custom stores on the platform that runs more of the web's commerce than anyone: variable products, subscriptions, checkout flows, gateways.
Oxygen Builder
The visual builder we use when a project stays on WordPress, and why it produces clean semantic HTML instead of shortcode soup.
HTML, CSS & JavaScript
How we use HTML, CSS, JavaScript, and browser APIs for custom websites.
W3C Standards
What compliance means here: markup that validates, carries meaning, and holds up under every browser, crawler and screen reader that touches it.
cPanel/WHM & CSF
Server administration and firewall hardening — how the machine a site lives on gets locked down, not just how the site gets built.
Open-source components
from Koda
Six Python components lifted straight out of Koda, our own production AI agent, and released under MIT. A credential vault, a verification engine that refuses to report success it cannot prove, deny-by-default permissions, workflows, a model router and tiered memory — 254 passing tests between them, one line of pip each. No signup, no telemetry, no strings.
Finding the Problem Is Free.
Fixing It Is the Job
Use the results to identify what needs a closer look. We can review crawler access, structured data, and page content, then help with the changes.