Humanneeds a URL + questions
AI Chatbot Quality Audit
Send the page your assistant lives on and the questions you care about. A person tests it and returns a scorecard with the raw answers.
Tools
Tools for deciding, not demos.
36 free tools that answer a question you actually have: is my current assistant any good, is my site ready for one, what would it save, what would it cost, and what should it be told to do. Nothing here pretends to scan your site automatically, because nothing here does.
- No account neededNothing to sign up for
- Public content onlyNo sensitive data required
- A person runs the teardownNot an automated score
Everything here runs in your browser
37 tools
Two are run by a person. The rest compute in the page as you type, send nothing anywhere, and work whether or not you ever become a customer.
37 shown
Humanneeds ten minutes
Website Assistant Readiness Checker
Six checks you can do yourself that decide whether an assistant will have anything to answer from.
Calculatorneeds your ticket volume
Support Deflection Calculator
Tickets and cost per ticket into a defensible saving. Defaults to 20% deflection, not the 60% vendors quote.
Calculatorneeds your traffic
Chatbot Cost Estimator
Traffic into conversations, credits, a plan and a price, with the headroom you would have left.
Calculatorneeds page count
RAG Cost Estimator
Embeddings, storage, re-indexing and inference per month if you build retrieval yourself.
Calculatorneeds conversation volume
Human Handoff Estimator
How many conversations reach a person, and whether one person can cover the busiest hour.
Calculatorneeds your page count
Knowledge Base Sizing
How many pages are worth indexing, and which ones make answers worse by being there.
Calculatorneeds cost and leads
Assistant Payback Calculator
Break-even month counting deflection, leads and the setup hours most calculators leave out.
Calculatorneeds tickets, leads and cost
AI Chatbot ROI Calculator
Twelve-month net return with deflection savings and lead value counted separately, and setup included.
Calculatorneeds six answers
Build vs Buy Scorecard
Whether to build retrieval yourself, scored on evaluation and maintenance rather than infrastructure cost.
Analyserneeds your robots.txt
AI Crawler Access Auditor
Which of fourteen AI crawlers you allow, block, or catch by accident with a wildcard.
Analyserneeds your llms.txt
llms.txt Validator
Seven structural checks against the spec, and what to add where it fails.
Analyserneeds a page of text
Answerability Score
Whether your paragraphs still make sense once retrieval pulls them out alone.
Analyserneeds a page of text
Chunkability Checker
Where your page splits at a given chunk size, and which chunks arrive without context.
Analyserneeds an answer + its source
Citation Checker
Paste an answer and the page it cited. Find out whether the page actually says that.
Analyserneeds your JSON-LD
Schema Coverage Auditor
Which structured data types answer engines read, and which required fields you are missing.
Analyserneeds a page of text
Stale Content Scanner
The claims on a page that expire: years, prices, counts and roadmap language.
Analyserneeds questions + your FAQ
FAQ Gap Finder
Which real questions have no answer anywhere in your content.
Analyserneeds any text
Token Counter
Tokens, cost and context-window fit for anything you paste.
Analyserneeds your transcripts
Chat Transcript Analyser
Refusal rate, repeated topics, and where visitors asked for a person.
Analyserneeds ticket subjects
Support Ticket Theme Finder
Tickets grouped into themes by volume, so you write the page that removes the most work.
Test kitneeds your subject matter
Hallucination Probe Kit
Questions your site provably cannot answer, so you can see whether it refuses or invents.
Test kitneeds your sector
Refusal Boundary Test Kit
Both directions: what it wrongly answers, and what it wrongly refuses.
Test kitneeds a widget you own
Prompt Injection Test Kit
Four attack classes including the one most teams miss, hidden in indexed content.
Test kitneeds vendors + questions
Assistant Comparison Sheet
Your questions, applied identically to every vendor you are considering.
Test kitneeds your sitemap
Test Question Generator
A regression set from your own pages, with the edge cases assistants fail.
Generatorneeds a few choices
System Prompt Generator
Grounding, refusal and handoff rules you can paste into any vendor, including a competitor.
Generatorneeds your sector
Refusal Policy Generator
What the assistant must never answer, how it declines, and what it offers instead.
Generatorneeds tone + cover
Chat Widget Copy Generator
All nine strings a widget needs, including the failure ones, in one voice.
Generatorneeds your cover model
Escalation Runbook Generator
What triggers a handoff, what the person receives, and what the visitor was promised.
Generatorneeds a few choices
Assistant Persona Builder
A voice spec rather than a character. Covers how it sounds when it does not know.
Generatorneeds your Q&A pairs
FAQ Schema Generator
Plain questions and answers into valid FAQPage JSON-LD, correctly escaped.
Generatorneeds your sitemap
llms.txt Generator
A structured llms.txt draft, grouped into sections, ready to fill in.
Generatorneeds paths to protect
robots.txt Generator
Explicit rules for every AI crawler, with training and answering as separate decisions.
Generatorneeds your jurisdiction
AI Disclosure Generator
Wording and placement for telling visitors they are talking to software.
Generatorneeds which side you are on
Data Boundary Clause Generator
Training exclusion, retention, subprocessors and deletion, as contract language.
Generatorneeds your risk level
AI Vendor Security Questionnaire
The questions worth asking an AI vendor, and what a bad answer looks like.
Nothing matches that. Try a shorter word, or clear the filter.
What these are not
STATED PLAINLY
Neither tool is an automated scanner. The audit is a request that a person runs against your live site. The readiness checker is a checklist you complete yourself. A tool that implies a scanner and delivers a form wastes your time and ours, so both say which they are before you start.
Launch
Start with the one that matches your situation
If you already run an assistant, the audit is more useful. If you do not, the readiness check tells you whether it is worth starting.