✳AgAgentHub
Use cases›existing agent
OWNER-REVIEWED USE CASE

Krishi Ai for Routing Farmer Queries to Structured Recommendations

[1]

Published 11/1/2026 · Project source ↗

Krishi Ai routes farmer queries to specialist agents, grounds answers in tools and RAG, and returns a structured recommendation. [1]

Verification: README documented; not independently verified. These source claims are documented, not independently tested results.

The agricultural problem

A farmer query may call for specialist routing and tool-grounded answers before a structured recommendation is returned. [1]

What the AI agent does

An orchestrator selects specialist agents rather than relying on a single chatbot, then checks the evidence they use. [2]

Expected output

The documented output is a structured recommendation grounded in tools and RAG. [1]

Inputs and data

  • Provide a farmer query for routing to specialist agents. [1]

Workflow

  1. The orchestrator selects agents, runs typed tools, validates evidence and can revise its answer up to twice. [2]
  2. Inspect the structured recommendation returned from the tool- and RAG-grounded response. [1]

Tools and setup

  • From backend, run uv sync, then uv run uvicorn app.main:app --reload to start the backend. [3]

Costs

Unknown — the reviewed source does not verify a cost for this workflow.

Limitations

  • Live weather and market APIs remain out of scope until their keys are wired. [4]

Source evidence

Documentation-backed source excerpts, not independent tests of the product. Expand a source to inspect its exact recorded excerpt.

  1. Source [1]DocumentedMulti-agent decision support for farmers: route a query to specialist agents, ground answers in tools + RAG, and return a…
    Multi-agent decision support for farmers: route a query to specialist agents, ground answers in tools + RAG, and return a structured recommendation.
    Open original documentation ↗
  2. Source [2]DocumentedAn orchestrator selects a set of agents, runs typed tools, validates evidence, then revises (max 2 loops).
    Not a single chatbot. An orchestrator selects a set of agents (finance, market, weather, crop health, equipment, soil, knowledge), runs typed tools, validates evidence, then revises (max 2 loops).
    Open original documentation ↗
  3. Source [3]DocumentedRun bash # Backend cd backend uv sync uv run uvicorn app.main:app --reload # → http://127.0.0.1:8000/docs
    Run ```bash Backend cd backend uv sync uv run uvicorn app.main:app --reload → http://127.0.0.1:8000/docs
    Open original documentation ↗
  4. Source [4]DocumentedOut of scope / next * Image diagnosis (V3) * Larger labeled set (100–300) and tool-level eval * Live weather/market APIs…
    Out of scope / next • Image diagnosis (V3) • Larger labeled set (100–300) and tool-level eval • Live weather/market APIs (until keys are wired) • Voice / mobile
    Open original documentation ↗

Browse more use cases →