Tools
Three free utilities we built for our own work and left public, plus a short note on how they fit together. They answer one chain of questions: what do AI assistants currently say about you, can they reach your pages, and does your markup tell them what those pages mean. No account, no email, no upload.
AI Visibility Prompt Tester
Turn your product and market into five buyer-style questions, then run them yourself in the assistants you care about and record whether your brand appears and how it is described. It writes the prompts and nothing else.
Build test prompts →
AI Crawler Radar
Enter any domain and get its AI crawler access status, structured data coverage and a plain-language list of what blocks AI citation. It fetches only robots.txt, llms.txt and the homepage.
Run a scan →
Schema.org JSON-LD Generator
Generate Organization, WebSite, WebPage, BreadcrumbList, FAQPage and Service or Product JSON-LD from a form, then copy or download the blocks. It runs entirely in your browser.
Open the generator →
AI Policy File Builder
Build a robots.txt and llms.txt that state which AI crawlers may read your site, per crawler group and per training or retrieval role.
Build my policy files →
Answer Block Builder
Turn a question you answer well into an answer-first paragraph, a visible FAQ block and optional FAQPage markup.
Build the block →
AI Crawler Log Analyzer
Paste a sample of your server log and see which AI crawlers visited, how often, and what they fetched. Parsed in your browser.
Analyze the sample →
Step & Standards Block Builder
Turn a procedure into a step block, a standards-and-verification block and a pitfalls block, plus optional HowTo markup.
Build the blocks →
Content Workload Planner
Work out how many answer blocks a coverage plan needs and how many weeks it takes at your real pace. Arithmetic on your inputs, in your browser.
Build the plan →
Visual Metadata Builder
Build an ImageObject block, an alt-text pattern from your own words, and a readiness checklist for multimodal search. Never inspects your image.
Build the metadata →
AI Visibility Prompt Tester
Give it what you sell and your brand name, and it writes five questions a real buyer might ask an assistant. You run them yourself and keep the record.
What it produces
- Five prompts in a buyer voice: shortlist, direct comparison, what the assistant actually knows, purchase risk, and a late-stage check.
- Your brand and any competitors you name are placed inside the questions, so the replies are easier to compare over time.
- A plain-text download so you can keep the exact prompts you used, next to your notes.
What it cannot do
It does not query any assistant, measure citation share, produce a score, or store your inputs. Results vary by model, index, region and date, so a single run is a snapshot, not a measurement.
AI Crawler Radar
Enter any domain and the scanner fetches its robots.txt, llms.txt and homepage, then reports what an AI crawler would actually experience.
What it checks
- Access status for 17 named crawlers, grouped into search engines, retrieval agents (the ones that fetch a page to answer a question) and training crawlers.
- Whether a wildcard rule blocks unlisted agents, and whether robots.txt declares a Sitemap.
- Parseable JSON-LD on the homepage, the schema types it declares, and any syntax errors.
- Extraction basics: title, h1, meta description, canonical, and an llms.txt file.
What it cannot tell you
It fetches three public files once. It cannot see your server logs, your rankings, or whether any engine has cited you, and it keeps no record of the scan.
Schema.org JSON-LD Generator
Fill in a form and get the JSON-LD blocks to paste into your page head. The markup is assembled in your browser, not on our server.
What it generates
- Organization, with logo, description, sameAs profile links and a contact point.
- WebSite, to identify the site as an entity, plus WebPage and BreadcrumbList for the page you are marking up.
- FAQPage built from the question and answer pairs you enter, and Service or Product for the subject of the page.
What it cannot do
It produces syntactically valid markup. It cannot guarantee rich results, rankings or citations, and it will not write your descriptions for you. Vague input produces vague structured data.
AI Policy File Builder
Answer a few questions about search engines, retrieval agents and training crawlers, and it writes the two policy files for you.
What it produces
- A robots.txt with one group per crawler, grouped into search engines, AI retrieval agents and AI training crawlers, plus an optional Sitemap line.
- A Content-Signal line that records your search, ai-input and ai-train choices in one place.
- An llms.txt that mirrors the policy, so a reader and a machine get the same answer.
What it cannot do
It cannot make a crawler obey the file, and it cannot read your current robots.txt. robots.txt is a request that well-behaved crawlers follow, not a lock, and a scraper that ignores it also ignores your new file.
Answer Block Builder
Give it one question you answer well, your direct answer and a few supporting facts, and it formats the block an engine can extract.
What it produces
- An answer-first paragraph, for placement directly under the heading that asks the question.
- A visible FAQ block in details and summary markup, with your facts as bullets and your sources named.
- Optional FAQPage markup, with a reminder that it belongs only alongside the visible block.
What it cannot do
It cannot write your answer or invent facts to fill a gap, it cannot measure whether an engine cites the block, and it cannot guarantee rich results.
AI Crawler Log Analyzer
Paste a sample of your server log and it counts which AI crawlers actually visited, in your browser.
What it produces
- A per-engine hit table with each crawler's share of recognised hits and its most requested path.
- Counts for lines parsed and lines unread, plus the first and last timestamp in the sample.
- A short summary you can copy into your own notes.
What it cannot do
It cannot verify that a user agent is genuine, since identification strings can be faked. It cannot read your analytics, and it does not store or transmit the lines you paste.
Questions
Are the tools really free?
Yes. No account, no email and no paywall.
Do you store what I enter?
No. The generator and the prompt tester run entirely in your browser, and the scanner keeps no record of a scan.
How do I test whether AI engines recommend my brand?
Build a set of buyer-style questions for your market, run each one in a fresh conversation in the assistants you care about, and record whether your brand appears and how it is described. The Prompt Tester writes those questions from your inputs. It does not run them for you.
Can I just ask an AI assistant to check my robots.txt for me?
You can, and a browsing assistant may read that file. What a scan adds is repeatability: the same rules applied to 17 named crawlers, plus your structured data, with the same output each time. You can rerun it next month and compare what changed.
Why not paste the markup from a chat window instead of using the generator?
That works for a quick fix. The generator's advantage is consistency: one set of blocks built from your inputs, downloadable as a file you can keep, version and diff whenever your site changes.
Do I need a tool to write test prompts?
No. The Prompt Tester makes the prompts consistent and repeatable and gives you a text file to keep your dated notes beside. If you prefer, write the questions yourself.
Does any of this guarantee AI citations?
No, and nothing can. Access and clarity are necessary conditions, not guarantees. What improves the odds is specific, verifiable content on the page itself.
Why not just use a commercial audit tool?
Most audit suites are built around search rankings. These three check the conditions AI engines care about, and we use them on this site ourselves.
All three tools are maintained by the same team that publishes this site. They produce diagnostics, prompts and valid markup. They do not promise rankings, rich results or citations.