KissanAI builds domain-specific vertical AI models and agent harnesses for agriculture. Our work turns agricultural knowledge, field data, language, vision, and operational workflows into AI systems built for how farming actually works.
Agriculture feeds the world, yet remains one of the least digitized industries. General-purpose AI is not enough for this domain. Agricultural AI needs crop context, regional practice, weather and climate intelligence, product knowledge, field memory, and workflows that move from question answering to operational action.
KissanAI is building that foundation: agriculture-specific language and vision models, curated agricultural datasets, evaluation pipelines, and agentic systems that can support farmers, agronomists, enterprises, and institutions across different farming environments.
|
Dhenu is the agentic operating system for agriculture, built for smallholder contexts where language, accessibility, local crop practices, and trusted advisory workflows matter. It brings together vertical agricultural models, voice-first interaction, and production-ready AI agents for farmer advisory, field teams, agribusinesses, and public institutions. |
FieldFoundry is the AI command center and intelligence layer for modern large-scale agriculture. It keeps field history, scouting, monitoring, product evidence, recommendations, and grower follow-through connected in one working system for agronomy teams and agricultural enterprises. |
- Agriculture-specific language, vision, and reasoning models.
- Agent harnesses that connect models to real agricultural workflows.
- Data curation and evaluation systems for reliable agricultural AI.
- Voice, multilingual, and field-ready interfaces for knowledge access.
- Intelligence layers for smallholder and large-operation agriculture.
Public source repositories are being prepared. For now, KissanAI's public model and dataset work is available on Hugging Face.
- ThinkingDhenu1-CRSA-India-preview - agricultural reasoning model for India-focused crop and climate advisory research.
- Dhenu2-In-Llama3.2-1B-Instruct - compact agricultural instruct model for India agriculture workflows.
- Dhenu2-In-Llama3.1-8B-Instruct - larger agricultural instruct model for advisory and agent workflows.
- Dhenu-vision-lora-0.1 - agricultural vision-language LoRA for crop disease and field image understanding.
- ThinkingDhenu1-Extension-USA-research-preview - research preview for US extension-style agricultural knowledge workflows.
- Thinking-climate-100k - climate reasoning data.
- india-climate-qa-kb-v1 - India climate question-answering knowledge base.
- india-climate-qa-synth-v1 - synthetic India climate question-answering dataset.
- KissanAI: https://kissan.ai
- Dhenu: https://dhenu.ai
- FieldFoundry: https://fieldfoundry.ai
- Hugging Face: https://huggingface.co/KissanAI
For collaboration, research access, or enterprise deployments, reach KissanAI through the website.
