BASTI
On-premises AI assistant with chat, knowledge spaces, RAG, and optional speech features.
BASTI is an on-premises AI assistant for organizations that need to run LLM-based capabilities entirely inside their own infrastructure. It combines a web-based chat interface, document-backed knowledge spaces, retrieval-augmented generation, knowledge graph visualization, document ingestion, and optional speech features in a layered architecture built from a UI, a gateway, and Python services. The project is designed for data sovereignty, predictable operations, and practical reuse in technical or knowledge-heavy environments. BASTI runs on Linux with Docker Compose, expects OpenAI-compatible inference endpoints, and can be operated without cloud dependencies in restricted networks.
Software-Details
07/13/26
07/13/26
beta
- web
- linux
0.1.5
EUPL-1.2
Making software use visible!
Last updated: