InformlyInformly Docs
Documents

Uploading documents

Add documents to your Informly library by uploading a file, pasting plain text, or crawling a URL, and follow the processing pipeline to completion.

Uploading is how most documents get into Informly. The library accepts files you already have on disk, plain text you can paste straight in, and web pages Informly fetches for you. Whichever you pick, the document moves through the same processing pipeline before the AI can answer from it.

Open the uploader from the sidebar by going to Documents → Upload, or click the create button on the library page.

Three ways to add a document

Pick the method that matches where your content lives today.

Drag a file onto the upload card, or click to open a file picker. Best for PDFs, Word documents, and .txt files you already have. See the file limits below.

Paste plain text directly into the editor and give it a title. Best for short snippets, internal notes, or content from a system Informly can't crawl. No file is created on disk — the text becomes the document.

Hand Informly a public URL and it fetches the page, extracts the main text, and ingests it. Best for help center articles, blog posts, and product pages. See Web crawl for the full flow.

File limits

LimitValue
Max file size20 MB per file
Supported types.pdf, .doc, .docx, .txt
Files per uploadOne at a time

If you have a source larger than 20 MB, split it before uploading — see the tip near the end of this page.

Upload a file

Open the uploader

Go to Documents → Upload.

Drop the file in

Drag a PDF, Word document, or .txt file onto the drop zone, or click it to open a file picker. Informly checks the size and type before the upload starts.

Set visibility

Choose whether the document is public to the organization or private. You can change this later from the document detail page or in bulk.

Save

Click Upload. The document appears in the library with a Pending status badge.

Paste plain text

Switch to the text tab

On the upload page, switch to the Paste text tab.

Give it a title

The title is how the document appears in the library and in chat citations. Make it descriptive — "Refund policy 2026" beats "Notes".

Paste your content

Drop the text into the editor. Plain text works best; rich formatting is stripped during processing.

Save

Click Create. Processing starts immediately.

What happens after upload

Every document, regardless of source, runs through the same four-stage pipeline:

  1. Text extraction — Informly pulls the readable text out of the file.
  2. Chunking — the text is split into short passages of a few paragraphs each.
  3. Embedding — each chunk is converted into a vector representation.
  4. Indexing — the vectors are stored so the AI can retrieve them at query time.

The status badge updates in place: Pending → Processing → Completed. You can leave the page and come back — nothing is lost.

You don't need to wait on the page during processing. The badge updates live the next time you visit the library, and the document is ready to assign to a widget as soon as it hits Completed.

When processing fails

If the badge turns red, open the document detail page to see the reason. The most common causes:

CauseFix
Scanned PDF with no text layerRun the PDF through OCR first, then re-upload.
File over 20 MBSplit the file into smaller documents.
Password-protected fileRemove the password and re-upload.

You can delete the failed document and try again, or fix the source and use Reprocess — see Managing.

Tips for cleaner ingestion

A few habits make a noticeable difference to answer quality:

  • Split very large books into chapters. A 200-page manual chunks better as ten 20-page documents than one giant file.
  • Strip boilerplate. Repeating headers, footers, and navigation menus eat into the chunk's useful content.
  • Use descriptive titles. Titles appear in citations and help your team find documents in the library.
  • Pick the right visibility up front. It's easier than fixing it in bulk later.

What's next

On this page