Document Conversion (MarkItDown)
Pando integrates the MarkItDown library to convert rich document formats into Markdown for Knowledge Base ingestion and RAG (Retrieval-Augmented Generation).
Supported Formats
| Format | Extensions |
|---|---|
.pdf | |
| Word | .docx |
| Excel | .xlsx, .xls |
| PowerPoint | .pptx |
| Web | .html, .htm |
| Data | .csv |
| E-book | .epub |
| Notebook | .ipynb |
| Feed | .rss, .atom |
| Markup | .xml, .json, .jsonl |
| Archive | .zip |
| Text | .txt, .md, .markdown |
CLI Usage
Convert a single file:
pando convert document.pdf
pando convert presentation.pptxKB Integration
When syncing a directory to the Knowledge Base, Pando automatically converts supported documents:
pando kb import /path/to/documentsDocuments in supported formats (.pdf, .docx, .xlsx, .pptx, etc.) are automatically converted to Markdown before indexing.
Configuration
No explicit configuration needed. Conversion is enabled automatically when the converter is available.
The conversion engine is lazily initialized to avoid startup costs from PDF processing.
How It Works
- KB sync detects a non-Markdown file
- Checks if the extension is in the supported list
- Converts to Markdown using the appropriate parser
- Indexes the Markdown content with front-matter metadata
- Original file is not modified
Place documents you want indexed in a KB-watched directory. They’ll be automatically converted and searchable. This is ideal for project documentation, specifications, and design docs stored as PDFs or Office documents.