MORAWA-dev/dakikobo

French-language RAG advisor for smallholder farmers in Burkina Faso. Groq Llama 3.3, ChromaDB, Gemini Vision leaf screening, voice in/out, CI eval harness. Live demo on HF Spaces. --- title: DakiKobo sdk: docker app_port: 7860 suggested_hardware: cpu-basic startup_duration_timeout: 1h preload_from_hub: - sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2 --- # DakiKobo — AI Agricultural Advisor for Burkina Faso 🌾 DakiKobo is a French-language AI assistant for smallholder farmers in Burkina Faso. It uses a **Retrieval-Augmented Generation (RAG)** pipeline grounded in agricultural reference documents (FAO, AGRA, WFP and technical guides for the Sahel and Sudanian Savanna zones) so answers stay accurate and source-backed rather than invented. The focus crops are **mil (millet), sorgho (sorghum), maïs (maize), niébé (cowpea) and arachide (groundnut)**. All output — answers, UI labels and voice — is in French, and the interface is mobile-first for use on phones. ## Essayer la démo (60 s) Live Space: **kimcomehome-dakikobo.hf.spa… 1. **Contexte parcelle** (optionnel) — culture, stade, lieu ; activez **Français simple** pour des phrases plus claires. 2. **Question texte** — ex. « Quand semer le mil ? » → carte *Conseil agricole* + sources. 3. **Engrais** — ex. « Dose d'engrais pour le sorgho » → doses **déterministes** (pas inventées par le LLM) + disclaimer agent. 4. **Photo de feuille** — dépistage prudent (*pas un diagnostic*). 5. **Sources & limites** — preuves, météo/sol indicatifs, confirmation terrain obligatoire. Script détaillé : `DEMO_SCRIPT.md`. Collecte de données (plus tard, hors code) : `Data/reviews/DATA_COLLECTION_TASKS.md`. Continuité agent : `SESSION.md`. **Ce que DakiKobo n'est pas :** un oracle de rendement, un diagnostic officiel, ou un remplacement de l'agent agricole. --- ## Features - **Grounded French answers** — RAG over a local document corpus; off-topic questions fall back to an honest "je ne sais pas" instead of hallucinating. - **Source citations** — each answer shows which document(s) it was drawn from. - **Fast inference** — Groq-hosted `openai/gpt-oss-120b`, with reasoning tokens hidd …