This is an interactive demo. In the real product, you upload your own documents and the AI learns from them using RAG (Retrieval-Augmented Generation).

Knowledge Base

Upload Documents
11 formats supported

Upload files and SpiritBox will chunk, embed, and index them into your private knowledge base. Ask questions in chat and your AI will cite sources from your documents.

Drag & drop documents here
PDF DOCX TXT MD CSV XLSX JSON JSONL HTML RTF CODE
Indexed Documents
6 documents
DocumentTypeChunksIndexedStatus
Q4_Financial_Report.pdf PDF 47 chunks 2 days ago Indexed
Product_Requirements.docx DOCX 23 chunks 3 days ago Indexed
Customer_Feedback_2026.csv CSV 156 chunks 5 days ago Indexed
API_Documentation.md Markdown 34 chunks 1 week ago Indexed
Budget_Projections.xlsx Excel 12 chunks 1 week ago Indexed
Meeting_Notes_Feb.txt Text 8 chunks 2 weeks ago Indexed
How Knowledge Base Works
1. Upload Drop in your files
2. Chunk Smart text splitting
3. Embed pgvector indexing
4. Retrieve AI cites your docs

All processing happens locally. Your documents are chunked, embedded via pgvector, and stored in PostgreSQL on your machine. When you ask a question in chat, relevant chunks are retrieved and injected into the LLM's context — giving it access to your private knowledge without ever sending data to the cloud.

Knowledge Stats
Documents indexed6
Total chunks280
Embedding modelall-MiniLM-L6-v2
Storage used42 MB
Databasepgvector
Memory Graph (Neo4j)
Entities42
Relationships67
Approved facts128
Rejected facts3
DatabaseNeo4j
Consent-gated: Every fact requires your explicit approval before storage.
Best Practices
  • Use clean, well-formatted documents for best results
  • PDF tables are extracted with pdfplumber for accuracy
  • CSV/Excel files are chunked row-by-row with headers
  • Markdown preserves code blocks and structure
  • Large files are automatically split into optimal chunks
  • Re-upload a file to update its index
Back to Overview