Reference
Accepted files and knowledge limits
Which formats you can upload, what cannot be read, the limits on size, pages and sources, and what each error message on a source means.
- Reviewed on
- 10 min read
In this article
You can upload PDF, Word (.docx), text (.txt, .md) and CSV files, usually up to 10 MB each, and import public web pages of up to 2 MB. The number of sources depends on your plan. Here is each limit, where to see yours, and what each error message means.
Files you can upload
| Format | Extension | Conditions |
|---|---|---|
.pdf | With text that can be selected | |
| Word | .docx | The old .doc format is not accepted |
| Text | .txt, .md | Plain text, saved as UTF-8 |
| CSV | .csv | Saved as UTF-8 |
The app checks the extension and the content of the file as well, so changing the extension of another kind of file does not work. The text of a document is read: any images inside it are not.
What cannot be read
- Other formats: old Word (
.doc), Excel spreadsheets, presentations, images. Save them as PDF or.docx; save a spreadsheet as CSV. - Scanned PDFs, or PDFs made of photos of pages: they have no selectable text. Use the original document or type the content in as Text.
- Password-protected or damaged files. Remove the password or export the document again.
- Text inside images: a photographed menu, a poster with your opening hours. Type it in as Text or Business info.
- Web pages that are not public: an intranet or a password-protected area. Copy the content and add it as Text.
- Pages that only render with JavaScript. Copy the content and add it as Text.
- Addresses that lead to a file, such as a PDF hosted on your website. Download it and upload it as a document.
- The menus, footer and forms of a web page. If a fact is only there, add it as Business info.
Limits
| Limit | Value | Where you see it |
|---|---|---|
| Size of a file | Usually 10 MB; never more than 25 MB | The Upload documents window shows the usual limit. If your account has a different one, the refused file states the real maximum: “The file is larger than the … maximum” |
| Size of a web page | 2 MB | “The page is larger than the 2 MB limit” |
| Pages in a crawl | As many as you ask for, between 1 and 200, with a platform cap that is usually 25 | In the list, next to the source: “18 of 25 pages”. In its details, the Crawl figure |
| Text in a source | Usually 400,000 characters | For a text or an FAQ source, the notice “The content is too long…” states your limit. A website or a document is read up to that limit and the rest is dropped: look at the Characters figure in the details |
| Questions in an FAQ source | 200 | The counter in the window, for example “3 of 200” |
| Length of a question and of its answer | 300 and 3,000 characters | “Use at most 300 characters in the question” |
| Name of a source | 200 characters | “Use at most 200 characters in the name” |
| Number of sources | Depending on your plan | Under the Sources heading, the line “Sources in the … plan”. Under Billing, the Sources figure |
| Re-indexes | 10 per hour per person | “You have reached the limit of 10 re-indexes per hour. Try again later” |
In addition, if you add, edit or test many times in a row, the app asks you to pause: “Too many requests in a row. Wait a moment and try again”.
Sources depending on your plan
The source limit is for the whole account: it counts the sources of all your agents. They count like this:
- A website is one source, whether it is one page or a crawl of many.
- Each file you upload is one source.
- An FAQ source is one source, with one question or with 200.
- Each text and each block of business info is one source.
When you delete a source, its place is free at once. This is what each plan includes:
Basic
- Price per month
- €99 + VAT
- Credits per month
- 3,000
- Agents
- 1
- Knowledge sources
- 30
- Team members
- 2
- Languages per agent
- 1
- Channels
- Web chat
- Advanced AI models
- No
- Extra credit purchases
- Yes
Standard
- Price per month
- €199 + VAT
- Credits per month
- 8,000
- Agents
- 3
- Knowledge sources
- 150
- Team members
- 5
- Languages per agent
- 3
- Channels
- Web chat, WhatsApp, Instagram
- Advanced AI models
- No
- Extra credit purchases
- Yes
Premium
- Price per month
- €299 + VAT
- Credits per month
- 20,000
- Agents
- 10
- Knowledge sources
- 500
- Team members
- 15
- Languages per agent
- Unlimited
- Channels
- Web chat, WhatsApp, Instagram
- Advanced AI models
- Yes
- Extra credit purchases
- Yes
How a website is crawled
- It starts at the page you give, which is always read, and follows the links to other pages of the same site.
www.yourcompany.comandyourcompany.comare the same site; a subdomain such asblog.yourcompany.comis a different one and needs its own source. - It does not enter the pages your website forbids to robots in its
robots.txtfile, and it does not follow links to files such as a PDF or an image. - It skips repeated pages and pages with hardly any text.
- It stops when it reaches the maximum number of pages or the text limit of the source, or when it has been running for several minutes.
Add the pages that were left out one by one with This page only: see add a website.
Messages when uploading a file
They appear in the Upload documents window, under the name of the file, which is marked “Not uploaded”.
| Message | What it means | What to do |
|---|---|---|
| “Unsupported format. Upload a PDF, DOCX, TXT, MD or CSV file” | The file is of another kind | Save it in one of those formats |
| “The file is empty” | It has no content | Check that you chose the right file |
| “The file is larger than the … maximum” | It is bigger than your account accepts; the message says how much that is | Split it into several files or make it smaller. If it is a short text, paste it in as Text |
| “The file is larger than the maximum your account accepts” | The same, without the figure | The same |
| “The file is not a valid PDF” | Its extension is .pdf, but its content is not a PDF | Export it as a PDF again from the original program |
| “The file is not a valid Word (.docx) document” | It is damaged or not a real .docx, such as an old .doc with the extension changed | Open it in your word processor and save it again as .docx |
| “The text is not UTF-8. Save it with that encoding and upload it again” | The text or CSV file uses a different character encoding | When you save or export it, choose UTF-8 encoding |
| “The file is not plain text. Save it as UTF-8 text and upload it again” | Its extension is a text one, but it contains data that is not text | The same |
| “The document does not contain any text” | The Word document has no text to read | Add the content as Text |
| “We could not read this file” | The check of the file failed or took too long | Upload it again. If it happens again, try a smaller file |
| “Not uploaded: you have reached the source limit of your plan” | There are no free sources left in your plan | Delete sources you do not use or change plan |
Messages on a failed source
When a source is Failed, the reason appears under its name in the list and, in its details, under “It could not be indexed”. Once you have fixed the cause, press Re-index: see keep knowledge up to date.
On a website
| Message | What it means | What to do |
|---|---|---|
| “The web address is not valid” | The saved address is not a correct web address | Delete the source and add it again with the full address |
| “This address cannot be imported because it points to a private or internal network” | It does not lead to a public website | Use the public address of the page or copy the content in as Text |
| “The website could not be reached. Check the address and index again” | The domain does not exist or your website was not responding | Open the address in your browser. If it is misspelt, add the source again; if it works, press Re-index |
| “The website took too long to respond” | Your website did not answer in time | Press Re-index later. If it keeps happening, mention it to whoever looks after your website |
| “The website answered with an error (HTTP …)” | Your website returned an error instead of the page. The number says which: 404, the page does not exist; 401 or 403, it asks for permission or blocks automated readers; 500 or above, your website is failing | Check the address. If your website blocks access, copy the content in as Text |
| “The address is not a web page. Only HTML pages and plain text can be imported” | It leads to a file, such as a PDF or an image | Download the file and upload it as a document |
| “The page is larger than the 2 MB limit” | The page is too big | Copy the part that matters and add it as Text |
| “The address redirects too many times” | It jumps from one address to another without reaching the page | Open it in your browser and use the final address you land on |
| “No readable text was found on the page. Pages that only render with JavaScript cannot be imported” | The page carries no text until it loads in a browser | Copy the content and add it as Text |
On a document
| Message | What it means | What to do |
|---|---|---|
| “The file could not be read. It may be damaged or password-protected” | The file cannot be opened | Remove the password or export it again, upload it once more and delete the source that failed |
| “The file contains no selectable text. Scanned documents are not supported” | It is a PDF made of images of pages | Use the original document or type the content in as Text |
| “The file took too long to process” | It is too large or too complex | Press Re-index. If it happens again, split it into smaller documents |
| “The original file is no longer available. Upload it again” | The file you uploaded is no longer stored | Upload it again and delete this source |
On any source
| Message | What it means | What to do |
|---|---|---|
| “This source has no text to index” | There is no content to make passages from | Edit it to add content, or delete it |
| “The AI service is unavailable right now. Index again later” | A temporary fault in the service; it has nothing to do with your account or the source | Press Re-index after a while. If nothing changes, write to us from Support |
| “Indexing was interrupted. Try again” | The process was cut off half-way | Press Re-index |
| “Indexing failed because of an internal error. Try again” | An error on iAgentify’s side | Press Re-index. If it happens again, write to us from Support with the name of the source |
The Waiting status is not an error and has nothing to do with your account: the source is saved and is indexed automatically as soon as the AI service is available again. There is nothing for you to do; if it stays like this, write to us from Support.
Related articles
- Add knowledge: your website, documents, texts and FAQsHow to add each kind of source to your agent (website, documents, FAQs, text and business info), which one to choose and what happens afterwards.
- Why your agent says it does not know, and how to fix itYour agent only answers with what is in its knowledge. If it says it does not have that information, here is how to find the gap and close it.
- Keep knowledge up to dateWhat to do when a price, your opening hours or your website changes: edit a source, re-index it, replace a document, delete it and check the result.
- Plans and credits: how they workA monthly plan includes credits and some limits, and you can add extra credits. What uses credits, what does not, where to see your balance and what happens at a limit.