RAG Knowledge Base

Description

Retrieval-Augmented Generation (RAG) system with semantic search using vector embeddings. DOCUMENTS: upload and rag PDF, DOCX/DOC, PPTX (with slide text, speaker notes, and optional visual analysis of slides with few words), Excel (XLSX/XLS), CSV, JSON, HTML/HTM, TXT, and Markdown. PDFs that are tabulated price catalogs are automatically detected and converted into a searchable SQLite table (tested on catalogs of hundreds of pages and thousands of references). MULTIMEDIA: upload images (PNG/JPG/GIF/WEBP), audio (MP3/WAV/OGG/M4A/AAC/FLAC), and video (MP4/WEBM/MOV). A multimodal model catalogs each file (description + tags) and indexes it for semantic search. The gallery displays video posters (generated with ffmpeg), playback with seeking (HTTP Range), and the chat embeds playable images, audio, and video from the retrieved sources. DATABASES: upload SQLite files (.db/.sqlite/.sqlite3) or SQL dumps (.sql, which are imported into a searchable database). The query engine is read-only (only SELECT/WITH). In the chat, the AI automatically generates SQL from your questions using the schema and a sample of data from your databases (supports multiple databases at once). ORGANIZATION: group knowledge into collections with subfolders of up to 3 levels. A reusable tree selector allows filtering by collection; selecting a folder automatically includes its subfolders, and disabling a folder disables its entire subtree for search. COLLABORATION (optional): with KB Workspaces enabled, each user can create named spaces and share specific collections of their KB with others via email (invitation with token + acceptance). Three roles: VIEWER (consults and uses shared collections in chat/RAG), EDITOR (also uploads, edits, and deletes documents, which remain under the identity of the KB owner), and AUDITOR (also sees the search history of the space, attributed to who searched and never exposing the owner's private searches). A member can leave a space at any time. A shared space can also be used as a KB source for an Agent or OmniChat bot, consulting it with the identity and scope of the owner's collections. Dual-switch governance and legacy-safe: the super administrator enables the feature at the platform level, and on that, each owner opts in for their KB; with the feature disabled (by default), each KB is 100% personal, and all functionality remains hidden and inert. INTEGRATION: the AI automatically uses your knowledge in the chat (with toggle), and is also available for Agents and OmniChat with configurable privacy scope. It also includes deterministic data analysis (exact calculations on Excel/CSV/JSON/SQLite, without LLM) and usage statistics.