Import from your website
Enter your website address and OMAI finds the useful pages and shows them to you before anything is added. What is taken, what is left out and why, how to refresh, and what happens to a page you remove from your site.
- Who
- Owners and admins
- Plan
- All plans
- Role
- Owner or admin
- Takes
- A few minutes to scan; a minute or two more to process
Before you start
- A public website with an https address
- Confirmation that you own the site, or are authorised to import its content
If your business has a website, it already holds most of what your assistant needs to know: what you do, what it costs, when you are open, and what happens on a cancellation. Import from your website turns those pages into knowledge sources — but no page is added until you have seen it and approved it.
How it works
Step 1 Enter your website address
You should see: On the Knowledge screen, in the Import from your website card. If you already entered your website in business settings, it is filled in for you.
Step 2 Wait for the scan
You should see: This can take a few minutes. You can stay on the page — it updates itself.
Step 3 Choose what to add
You should see: You get a list of the pages found. The recommended ones are already ticked, and every page says why it was recommended or why it was left out. You can open any page to check it.
Step 4 Confirm the site is yours
You should see: Tick the box confirming you own the website, or are authorised to import its content. Nothing can be added without it.
Step 5 Add to knowledge
You should see: The pages you chose appear in your knowledge sources and start processing. Once they reach Ready, your assistant can answer from them.
What is taken and what is not
The aim is business content, not every page on the site. Navigation menus, headers, footers, forms and code are removed before the content is saved, so your assistant never learns your menu.
| Page type | Why |
|---|---|
| Login, sign-up, account | Nothing there helps a customer asking a question |
| Cart, checkout, order confirmation | Action pages, not information pages |
| Search results | The content changes with whatever was searched for |
| Archives: tags, categories, author pages, date archives | Lists of links, not content |
| Calendars and month-by-month events | There is no limit to how many such pages exist |
| The same page in another language | The copy in your business language is kept |
| Almost-empty pages | Under 200 characters is not enough to answer from |
| Product pages, if you have an OMAI catalog | The catalog is a more reliable source for products |
Refreshing
Websites change, and your knowledge should change with them. Each imported website has a Refresh button that scans it again and shows you what changed.
| Page state | What happens |
|---|---|
| Unchanged | Nothing — no needless reprocessing |
| Content changed | A new version is built. The old one keeps answering until the new one is ready |
| New page on the site | Shown for approval, never added on its own |
| Removed from the site | Disabled only after two consecutive checks find it gone |
| We could not read the page | Nothing happens. A temporary fault on your site never deletes knowledge |
Limits
- Up to 50 pages per scan, up to two levels deep from the home page
- Up to 3 scans per hour and 10 per day per business, and one scan at a time
- HTML and text pages only. PDFs linked from your site are not scanned — upload them through Upload a document
- Imported pages count towards your plan storage allowance, like any document
What OMAI does not do
What wins over what
Information from your website sits low on the reliability scale — below documents you uploaded, and far below what you entered in settings. If your site shows an old price and your services list an updated one, the assistant answers with the one you entered. That is deliberate: you updated it, and perhaps not the website.
I entered an address and no pages were found. Why?
Usually one of two things: the site is built entirely in JavaScript so the pages carry no readable text, or every page is too short. In that case upload the information as a file, or type it into services and questions.
My website blocked the scan. What now?
Your site’s robots.txt blocks automated crawlers and we honour it. You can change it on your site, or simply save the important pages as files and upload them.
I was told the domain belongs to another account.
That website has already been imported into another OMAI account. To continue you will need to prove control of the site — a tag on the home page, a file at /.well-known/, or a DNS record. Contact us and we will walk you through it.
Does importing from my website change the services and prices I entered?
No. Website import creates knowledge sources only and never touches the fields you entered in settings. Filling those in is a separate action — AI import, with its own approval screen.