ZettaBrain RAG

ZettaBrain RAG is a self-hosted document assistant for individuals and small teams working with policies, contracts and internal manuals. It helps them find answers and draft documents from files they already keep on their own infrastructure.

• Answers with citations: Ask questions in plain language and check the source document, page and passages behind each answer.
• Existing file storage: Read PDF, DOCX, TXT and Markdown files from local disk, NFS, SMB, S3-compatible storage or OneDrive while leaving the originals in place.
• Keyword and semantic search: Find exact names, codes and phrases alongside passages with related meaning, with re-ranking to select relevant material.
• Document drafts: Built-in generation skills use retrieved content to produce SOPs, compliance audits, RFP responses, contract summaries and runbooks.

You can access the assistant through an HTTPS browser interface or an interactive command-line interface. Ollama runs the language and embedding models locally, with ChromaDB storing embeddings on disk. Supported model families include Llama, Mistral and Qwen. Local operation needs no cloud account or API key and can work without outbound internet access.

Optional Claude, Gemini or OpenAI connections send queries and retrieved passages to the selected provider. ZettaBrain RAG runs on Linux, macOS and Windows via WSL2, with CPU-only operation or NVIDIA, AMD and Apple Silicon GPU acceleration. The separate commercial, self-hosted ZettaBrain Teams product provides per-team document isolation, Active Directory sign-on and audit logging.