Library

Upload documents and ask AI questions about them. Build your personal knowledge library that OpenLoomi can query.

---

Overview

The Library feature allows you to:

• Upload documents (PDF, Word, text, etc.)
• Ask AI questions about your documents
• Get precise answers with citations
• Build a searchable knowledge library

---

Supported File Types

---

How to Use

1. Upload Documents

1. Go to the Library page
2. Click Upload or drag-and-drop files
3. Wait for processing to complete

2. Ask Questions

Once uploaded, simply ask OpenLoomi:

• "What's in the document about XXX?"
• "Summarize the main points of this file"
• "Find information about [topic]"
• "What does section 3 say about [subject]?"

3. View Full Content

• Click on any document to view its full content
• Search within documents
• View metadata and upload date

---

Use Cases

Contract Review

Upload contracts and ask:

• "What are the payment terms?"
• "What's the termination clause?"
• "Are there any liability limitations?"

Report Analysis

Upload reports and ask:

• "What are the key findings?"
• "Show me the revenue numbers"
• "What recommendations are made?"

Research Papers

Upload papers and ask:

• "What methodology was used?"
• "What are the main conclusions?"
• "Summarize the abstract"

Meeting Notes

Upload meeting notes and ask:

• "What decisions were made?"
• "Who committed to what?"
• "What were the action items?"

---

Best Practices

File Organization

• Use clear, descriptive filenames
• Group related documents together
• Delete outdated files regularly

Questioning

• Be specific in your questions
• Ask one question at a time
• Use quotes for exact phrases

Security

• Only upload documents you have rights to share
• Sensitive documents stay within your account
• Delete documents when no longer needed

---

Tips

• Start with shorter documents to test
• PDFs with text (not images) work best
• Large documents may take longer to process
• You can upload multiple files at once

---

Verify

• The uploaded document appears in the Library page with the correct file type and size.
• A question about a specific phrase in the document returns a citation pointing back to the right chunk.
• Deleting the document removes it from search results without affecting the rest of Memory.

Troubleshooting

• Upload stalls. Confirm the file isn't above the size limit and that the page hasn't lost connection; re-upload if needed.
• Image-only PDFs return no answers. The RAG pipeline extracts text — scanned PDFs need an OCR pass first.
• Citations point at the wrong chunk. Re-upload the source if the original was edited; chunks aren't auto-refreshed.

Related

• Memory — the Knowledge Base lives inside Memory's rag_documents / rag_chunks tables
• Embedding Providers — choose cloud vs local embeddings for chunk vectors
• Chat — how Chat uses Library answers
• Connectors — auto-fetch documents from connected platforms
• Privacy & Security — Library data stays local