Your site already has the answer. Visitors just can’t find it.
Hooty reads your pages and puts an assistant and a smart search on your site, so a visitor gets the answer instead of a list of links — and every answer names the page it came from.
Free plan. No card, no sales call.
Search returns links. People came for an answer.
Someone with a question types it into your search box and gets a list of pages. Now they have to guess which one holds the answer, open it, and read until they reach the part that matters.
Often the answer really is on the site — spread over three pages, and nobody reads all three. So they write to you instead, and you answer the same question again.
Sites where the same questions arrive every day
Documentation and help centres
The answer is written down already, three pages deep. This is the difference between having documentation and people reading it.
Online shops
Delivery, returns, sizing, stock. Questions asked at the moment of buying, where waiting for an email reply means not buying.
SaaS and apps
Pre-sale questions your pricing page half-answers, and setup questions that turn into a support ticket on day one.
Services with a long FAQ
Clinics, schools, agencies, councils. Everything is on the site, spread over pages nobody reads end to end.
Crawl, index, answer
Three steps. Nothing starts until you add a site.
Crawl
HootyBot fetches your public pages, one at a time. It reads robots.txt first and follows it.
About HootyBotIndex
What it read becomes the material the assistant may use. Your pages, and nothing else.
Answer
A visitor asks in their own words. The answer is built from those pages and names the ones it used.
The three steps in detailA chat in the corner, or a search field in the page
Both read the same index of your site. Most people end up running both, because they answer different moods.
Chat widget
One line of JavaScript and a chat appears in the corner of your site. It answers from your pages and links to the ones it used.
Search module
A search field placed where you want it, returning ranked pages of your own site with a preview. No conversation — a click goes straight to the page.
See every conversation, and step into any of them
Read what your visitors asked and what they were told. When one needs a person, take it over and type yourself — the visitor stays in the same window. What the assistant could not answer is the clearest list you will get of what your site is missing.
It answers from your pages, or it says it cannot
Your content, not the web’s
The assistant works from the pages HootyBot read on your site. It is not a general chatbot wearing your logo.
Every answer is traceable
An answer names the pages it came from, so a visitor — or you — can open the source and check it.
Nothing invented
When your pages do not cover a question, that is the answer. The gap is not filled with something plausible.
You decide what is read
robots.txt decides which pages HootyBot may fetch. What it may not fetch never reaches the index.
HootyBot says who it is
If Hooty reads a site you run, the crawler is not an unknown line in your logs. Every request carries its name and an address you can write to.
- User-Agent
Mozilla/5.0 (compatible; HootyBot/1.0; +https://hooty.to/bot)- From
bot@hooty.to
- It fetches /robots.txt first and follows it, including * and $ patterns.
- One page at a time per site, at least half a second apart, honouring Crawl-delay.
- HTML pages only. It never submits a form, logs in, or touches a cart or a checkout.
- It does not use the pages it reads to train models.
- Two lines in robots.txt stop it. The instructions are in the crawler documentation.
Give your site something to answer with
Add a site, let HootyBot read it, and see what it can answer.