Managing documents
Search the Informly document library, run bulk actions, reprocess and archive documents, manage visibility, and download original files.
Once you've uploaded more than a few documents, the library stops being a list and becomes something you need to manage. This page covers the day-to-day operations: finding documents fast, changing things in bulk, refreshing content after edits, and getting files back out of Informly.
Everything here happens in the library at Documents.
Searching the library
The library supports two search modes that run side by side against the same content. Pick whichever matches the question you're asking.
| Mode | Best for |
|---|---|
| Full-text | "I remember the exact title or a unique phrase." Word-for-word matches. |
| Semantic | "I remember the topic but not the wording." Meaning-based matches. |
Both return the same document cards and let you click straight through to the detail page. Semantic search uses the same embeddings the AI uses at chat time, so if a document doesn't surface here it likely won't surface for the assistant either — a useful debugging signal.
Bulk actions
Multi-select makes it cheap to act on many documents at once.
Enter multi-select
Tick the checkbox on any document card. A selection bar appears at the top of the library.
Add more
Tick additional documents, or use the Select all option in the bar to grab the whole filtered view.
Run an action
Pick a bulk action from the bar:
- Bulk delete — permanently remove the selected documents.
- Bulk visibility — flip all selected documents to public or private at once.
Bulk delete is permanent — chunks and the original file are removed. If you only want to take a document out of circulation, archive it instead.
Reprocess a document
Reprocessing re-runs the full extract-chunk-embed-index pipeline against the document. Use it when:
- You've replaced the source file with a newer version.
- A crawled URL has changed and you want the latest copy.
- You changed an ingestion setting that affects how chunks are produced.
Open the document
Click the document in the library to open its detail page.
Click Reprocess
The status drops back to Processing and the badge updates as the pipeline runs.
Reprocessing replaces the existing chunks in place. The widget assignments, visibility, and metadata you set earlier are preserved.
Archive instead of delete
Archiving hides a document from the library and stops it appearing in chat responses, without deleting the file or its chunks.
| Action | Hidden from library | Removed from chat | Reversible | Original kept |
|---|---|---|---|---|
| Archive | Yes | Yes | Yes | Yes |
| Delete | Yes | Yes | No | No |
Archive is the right default when you're not sure. You can always unarchive later from the archived filter view.
Download the original
For any uploaded file, the document detail page exposes a Download action that returns the original file exactly as you uploaded it. Useful when:
- You need the source to edit and re-upload.
- A teammate without Informly access needs the file.
- You're migrating content somewhere else.
Crawled documents and pasted-text documents don't have a downloadable original — the document itself is the source.
Visibility
Visibility controls who in your organization can see the document in the library, separate from which widgets answer from it.
| Visibility | Who sees it in the library |
|---|---|
| Public | Every member of the organization. |
| Private | Only members with explicit access. |
Change visibility one document at a time from the detail page, or in bulk from the library multi-select bar.
Inside the document detail page
Each document has its own detail page with everything you need in one place:
- Metadata — title, type, size, status, visibility, source URL if crawled.
- Chunks — the passages the AI retrieves. Useful for checking that important content survived chunking.
- Generate summary — produce a short AI summary of the document on demand.
- Chat with document — a private sandbox where you can ask the AI questions against just this one document. Great for verifying answer quality before exposing the document to a widget.
The chat sandbox is the fastest way to spot a bad ingestion. If the assistant can't answer obvious questions here, it won't do better in a live widget.
What's next
Web crawl
Turn a public URL into an Informly document by fetching the page, extracting the main text, and indexing it alongside your uploaded files.
Data sources
Connect Informly to external systems like Slack, Salesforce, and Google Drive to sync documents into your library automatically and keep them up to date.