Bilingual health answers grounded in local knowledge
TengLaafi
A bilingual assistant for tropical health and medicinal plants, grounded in curated research and local knowledge.
Current state
Research prototype focused on grounded bilingual retrieval and locally relevant knowledge.
Case study
Problem
Generic health assistants rarely combine tropical-health research, medicinal-plant knowledge, and bilingual access for a local context.
Approach
TengLaafi grounds answers in a curated knowledge base instead of relying on an unconstrained language-model response.
Engineering
FastAPI exposes the service while LangChain, ChromaDB, and Mistral-7B support retrieval and answer generation.