Private Knowledge AI

Upload documents, index them locally in the Space runtime, and ask source-grounded questions.

v08 profile: hf_token
Embedding backend: sentence_transformers
Default LLM: llamaindex/vdr-2b-multi-v1 through Hugging Face Inference API.
Token status: True
Fallback: extractive RAG if token/API fails.

Upload files, then build a semantic RAG index. Supported: PDF, DOCX, TXT, MD, CSV, XLSX.

Documents

Reset index before ingest

Privacy note: HF API modunda soru ve retrieved context Hugging Face Inference API’ye gider. Public Space’e gizli belge yüklemeyin. Gerçek müşteri verisi için Private Space veya on-premise sürüm kullanın.