Word files convert to PDF for preview, but
.docx is preserved; Download returns the editable original, never the PDF. Handy for handing back source.Supported formats
| Category | Formats | Preview behaviour |
|---|---|---|
.pdf | Full preview: page navigation, zoom, highlight, citation jumps. | |
| Word documents | .docx, .doc | Auto-converted to PDF on upload, preview behaves like a PDF. |
| Excel | .xlsx, .xls, .csv | Cell-level preview. |
| Images | .png, .jpg, .jpeg, .tiff, .gif, .webp, .svg, .bmp | Inline image preview with zoom. |
| Email files | .eml, .msg | Plain-text preview of the email body, with attachments listed. |
| Text files | .txt, .md, .log | Plain-text preview. |
| Structured text | .csv, .xml, .json | Plain-text preview with format-aware highlighting. |
| HTML | .html | Rendered preview in a sandboxed frame. |
What’s the same across all formats
Whatever the format, four behaviours are consistent:| Behaviour | What it means |
|---|---|
| Click a citation, jump to the source | Whether the source is a PDF page, an image region, or a paragraph in an email, citations land you in the right place. |
| Tabs in the viewer | Multiple documents can be open as tabs along the top of the viewer. Switch between them without losing your scroll position. |
| Search and filtering by content | Global Search indexes the content of every supported format, not just PDFs. |
Format-specific notes
PDFs
PDFs
The richest preview. Page navigation, zoom, full-text highlighting, citation jumps, and translation controls (if your subscription includes them).
Word documents
Word documents
Word files are converted to PDF on upload, so the preview behaves identically to a PDF. The original
.docx is preserved; download to get the editable version.For editing Word documents in place with AI assistance, use the Word Add-in.Excel and CSV
Excel and CSV
Cell-level preview. Each sheet appears as a tab; cells are clickable and citations to specific cells are supported.
Images
Images
Rendered as images with zoom and pan. OCR-extracted text is indexed, so a scanned contract uploaded as an image is fully searchable. Citations into images point at the OCR-extracted passage.
Emails (.eml / .msg)
Emails (.eml / .msg)
Body rendered as text. The viewer shows From, To, Subject, Date, and the body. Attachments are listed; click them to open in their own viewer tab.Emails exported from the Outlook Add-in into a project are auto-converted to PDF for permanent indexing.
Text files
Text files
Plain text rendering with no formatting. Useful for transcripts, log files, and quick notes.
Structured text (CSV, XML, JSON)
Structured text (CSV, XML, JSON)
Rendered as plain text with light syntax awareness. For data analysis, prefer using Discovery against the file rather than reading it directly.
HTML
HTML
Rendered in a sandboxed frame so the page can’t navigate away or run scripts. Useful for archived web pages or saved articles.
Document tabs
The viewer keeps multiple documents open as tabs.
Switch tabs
Click any tab. The viewer switches to that document; your scroll position in each tab is preserved.
Citations across formats
A chat answer or a Review or a Discovery can cite any of these formats. The citation behaviour is the same:| Source type | Click a citation | Result |
|---|---|---|
| Citation badge in the answer. | Document opens in the viewer at the cited page, passage highlighted. | |
| Word document | Citation badge. | Converted-PDF preview opens at the cited paragraph. |
| Image | Citation badge. | Image opens with a marker on the cited region. |
| Citation badge. | Email opens, cited passage highlighted in the body. | |
| Text file | Citation badge. | Text file opens, cited passage highlighted. |
Downloading
Every preview has a Download button in the toolbar. Downloads return the original file (so a Word document downloads as.docx, not as the converted PDF).
Scanned PDFs and images are OCR’d in the background; text isn’t searchable until processing ends. No hits on a fresh upload? Wait a minute, re-query.
Tips
Upload originals, not converted versions
Upload originals, not converted versions
If you have the original
.docx, upload it. Libra converts it to PDF for preview but keeps the original, so the download is editable.Use OCR for scanned documents
Use OCR for scanned documents
Scanned PDFs and image files (TIFF, etc.) are OCR’d automatically. Wait until processing finishes before assuming search has nothing; the text appears in the index after OCR completes.
Email + attachments work together
Email + attachments work together
An exported email thread keeps its attachments as separate documents in the same folder. Citations to either work.
Use Discovery for data files
Use Discovery for data files
A folder full of Excel files or CSVs is a Discovery’s natural input. Don’t try to read them all in the viewer; extract the data into a Discovery instead.
Next steps
Folder structure
Organise the panel into folders.
Documents
Back to the Documents overview.