io.github.gabrielmahia/kilimo-mcp

kilimo-mcp

---

Compatible with claude-sonnet-5 (released 2026-06-30) — Anthropic's most agentic Sonnet yet. Runs multi-step tool chains end-to-end without stopping short. Install: pip install kilimo-mcp · Use with any MCP client.

---

Kenya precision agriculture via MCP — crop calendar, fertilizer, pest alerts, KALRO varieties

AI Architecture: Agricultural Domain RAG

Agricultural AI for East Africa requires domain-specific retrieval — generic models hallucinate crop varieties, soil types, and local market conditions.

Research basis:

• AgriIR (Springer 2026): configurable RAG framework for agricultural knowledge retrieval — validates domain-specific retrieval over monolithic agricultural models. Key finding: low computational cost with high accuracy when knowledge base is curated.
• RAG vs Fine-tuning on Agriculture (arXiv 2401.08406): RAG outperforms fine-tuning for dynamic agricultural knowledge (seasonal, location-specific, real-time market).
• DPGA 2026 focus: Domain-specific DPGs in agriculture and climate are a 2026 strategic priority for Co-Develop and DPGA members.

Knowledge base pairing:

• Kenya National Drought Management Authority (NDMA) data → drought → crop risk
• Kenya Meteorological Department seasonal forecasts → planting calendar
• East Africa agricultural historical data (1906-1925, HuggingFace: gmahia)

Coordination: Pairs with wapimaji-mcp (water), bima-mcp (insurance), soko-mcp (markets) via africa-coord-bus for full drought response cascade.

Install

pip install kilimo-mcp

What it does

6 MCP tools covering Kenya precision agriculture. 1st world equivalent: Climate Corp / Granular.

Tool	Description
crop_calendar	Kenya crop planting and harvesting calendar by region
fertilizer_guide	Fertilizer recommendations by crop and soil type
pest_disease_alert	Crop pest and disease identification and management
market_timing_guide	Best timing to sell produce based on price cycles
kalro_varieties	KALRO improved crop variety recommendations
input_cost_calculator	Kenya farm input cost calculator and comparison

Usage

# Run as standalone MCP server
kilimo-mcp

# Or add to Claude Desktop / any MCP client
# Add to your MCP config: {"command": "kilimo-mcp"}

Part of the Kenya Coordination Infrastructure Stack

This is one of 23 MCP servers covering the full coordination infrastructure of East Africa:

Economic: mpesa · mkopo · bima · soko · sifa · remit · kra · faida
Physical: wapimaji · nishati · usafiri · ardhi
Social: afya · afya-ya-akili · elimu · kazi · haki-ya-kazi · kilimo · jumuia
Civic: nyumba · habari · mazingira · civic-agent-kit

→ The Nairobi Stack
→ Full Portfolio

Trust Integrity

All data in this server is clearly labeled DEMO where synthetic. Verify all operational data with the relevant Kenyan government authority before use.

License

MIT © Gabriel Mahia | AI-KungFU | contact@aikungfu.dev

Decision infrastructure for East Africa

Part of the East Africa Coordination Stack

This MCP server is one of 32 tools in the Kenya coordination infrastructure. It connects to africa-coord-bus — the coordination event bus that routes signals between domains automatically.

When this server detects a threshold condition, the bus notifies:

• bima-mcp — parametric insurance evaluation
• kilimo-mcp — agricultural advisory
• afya-mcp — health surveillance activation
• county-mcp — county office alert

pip install africa-coord-bus

All servers: pypi.org/user/gmahia

IP & Collaboration

MIT licensed. Feedback via GitHub Issues only — pull requests are not accepted. Demo data is labeled DEMO and is not suitable for operational decisions. Full policy: docs/architecture/IP_POLICY.md. Security reports: see SECURITY.md.

Part of the East Africa coordination stack

• Install & run: pip install reli-cli && reli list — 33 MCP servers on the official MCP Registry under io.github.gabrielmahia
• Evaluate any model on Swahili agent tasks: kipimo · dataset · leaderboard
• Coordinate across servers: africa-coord-bus — offline-first event bus with a built-in Kenya routing table
• Datasets: huggingface.co/gmahia · Docs hub: nairobi-stack

Model-agnostic by design: closed APIs, open-weight models, and small distilled models are all first-class citizens.