KrishiBot helps farming teams make confident, data-informed crop decisions.
KrishiBot delivers practical agricultural support through disease analysis, guided advisory workflows, and conversational assistance tailored to real field operations.
What KrishiBot Delivers
The platform is designed to reduce uncertainty in field operations and support faster, clearer agricultural decisions across crop planning, protection, and production.
Disease Detection
Analyze crop and leaf photos to identify likely disease patterns and receive practical next-step guidance.
Advisory Guidance
Get clear recommendations for irrigation, nutrition, pest prevention, and seasonal crop care.
Conversational Support
Ask farming questions in natural language and receive concise, decision-focused responses.
Weather-Aware Decisions
Use weather context to plan field operations, reduce risk, and improve timing of critical actions.
AI Models
KrishiBot combines a computer vision classifier with a local language model to deliver accurate, private, and offline-capable agricultural support.
EfficientNet-B0 (Computer Vision)
- • Fine-tuned on PlantVillage dataset
- • 38 plant disease classes across 14 crop species
- • 98.2% validation accuracy
- • Runs locally, ~500MB RAM usage
Qwen2.5 3B (Language Model)
- • Natural language understanding and generation
- • Bilingual: English + Bangla
- • Served via Ollama locally
- • ~3.5GB RAM usage
How We Build Trust
KrishiBot is built around reliability, clarity, and ownership of data, so advisory support stays practical and dependable in real-world use.
Farmer First
Designed for practical use in real field conditions with clear language and actionable recommendations.
Private by Design
Farm data and images stay under your control, with a privacy-focused experience by default.
Reliable Offline Access
Built for regions with unstable connectivity so support remains available when it is needed most.
Operational Value
Teams using KrishiBot can standardize advisory workflows and respond more quickly to common crop health and planning challenges.