Welcome to the Document Embedding discussion forum! Document Embedding adds semantic search to Mayan EDMS using local language models. Purchase / details: https://mayan.mypalace.ca/apps/document-embedding/ WHAT DOCUMENT EMBEDDING DOES ============================= - Calculates a semantic fingerprint for each document - Find conceptually related documents even with different vocabulary - Works across multiple languages - 100% local — no data sent to any cloud service - Includes a "Similar documents" panel on each document page GETTING STARTED =============== 1. Install and add to INSTALLED_APPS 2. Run migrations 3. Configure the embedding model under Settings > Document Embedding 4. Run the initial indexing task: manage.py embed_all_documents 5. Use the semantic search bar or the "Similar" link on any document Common questions: - Which embedding model should I use? (recommended: all-MiniLM-L6-v2 for speed) - How long does initial indexing take? (depends on corpus size and hardware) - Does it work with scanned documents? (yes, via the OCR text) --- Bienvenue sur le forum Document Embedding ! Posez vos questions sur la recherche sémantique, les modèles d'embedding et la configuration. Achat / détails : https://mayan.mypalace.ca/apps/document-embedding/