Helping farmers sell at the right price
Built end-to-end in 4 hours, my first code project.
Role
Producer
Timeframe
4 hours
Team
Me + Claude
Platform
Discovery
I get real-time stock prices, instant notifications, personalized feeds. My father-in-law, a farmer in Karnataka, India, walks to local shops just to learn what his coffee is worth today.
Technology has transformed how I work and live. It has barely touched his.
He receives areca nut prices on his mobile once a week, in English, not his native language Kannada.
For coffee and black pepper prices, he physically visits local shops to ask.
Prices can swing from ₹70 to ₹300 in the same week, timing directly affects a farmer's entire year of yield.
No single source exists for all three commodities he grows.
Problem
Farmers sell at suboptimal prices due to information lag.
From a farmer's perspective
Weekly English updates aren't frequent enough for daily decisions.
Visiting shops takes time away from farming.
Kannada is his first language, English creates unnecessary friction.
Multiple commodities require multiple information sources.
A small delay in selling can affect an entire year's harvest revenue.
From a systemic perspective
Market information exists, it's just not accessible in the right language, at the right time, on the right channel.
The same data that reaches traders instantly takes days to reach farmers.
Indian farmers grow crops all year but depend on middlemen for price information. The data exists, it just doesn't reach them.
Impact
Farmers sell at suboptimal prices due to information lag.
Middlemen with better information make more money.
Time spent gathering prices reduces productive farming hours.
Language barriers add cognitive load to simple decisions.
Without daily price information, farmers sell at the wrong time and lose months of work.
Initiative
I could have designed an app. Instead, I chose the simplest path to value.
The problem was clear. The user was accessible. The constraint was my technical ability, I had never written code.
I worked with Claude Code to build end-to-end.
Automated daily delivery replaced manual information gathering.
No app to download. No new interface to learn. Just a WhatsApp message in Kannada, every morning at 10 AM.
Key Decisions
I made deliberate trade-offs to reach farmers where they already are.
Prioritized WhatsApp over a custom app: 98% of target users already use it daily.
Chose Kannada as the primary language: farmers should read prices in the language they think in.
Focused on three commodities only: areca nut, black pepper, coffee, the crops my father-in law grows.
Built stateless architecture: no database, fresh scrape daily, minimal maintenance.
Used free infrastructure: GitHub Actions + WhatsApp Cloud API = zero cost to build.
Shipped MVP first: validated with real users before adding features.
Built on free infrastructure: GitHub Actions triggers the scraper daily, fetches prices from three sources, and delivers via WhatsApp.
Research
Research combined direct observation and technical exploration.
Observed my father-in-law's existing price-checking behavior over multiple visits.
Identified the three commodity sources he cared about most.
Explored scraping feasibility for MAMCOS, CommodityMarketLive, and Kirehalli website.
Tested WhatsApp Cloud API capabilities and limitations.
Validated timing, 10:00 AM IST aligns with morning market decisions.
The three commodities: Areca nut, black pepper, and (Robusta) coffee, grown on the same farm, tracked from three different sources.
Design
Message structure for scanning
Designed the WhatsApp message for instant comprehension, date at top for context, commodity names as headers, min/max/avg prices for volatility awareness.
The message needed to be scannable in seconds, no mental translation required.
Language for trust
Kannada-first design eliminated the friction of English. Farmers read prices in the language they think in.
Timing for decisions
Messages arrive at 10:00 AM IST, after markets update, before selling decisions are made.
Consistent timing builds habit. Predictable delivery builds trust.
Channel for accessibility
WhatsApp was chosen deliberately:
Works on basic smartphones with low bandwidth
Messages are saved and searchable
No new app to learn
Already part of daily routine
Outcome
Speed: From idea to production in 4 hours. Research, architecture, coding, deployment, all in one session.
Cost: ₹0 infrastructure cost. Free to run, free to scale to more farmers.
Reliability: 100% uptime via GitHub Actions SLA. Stateless architecture means no maintenance burden.
Learning: Day 1 of GitHub. First repository, first pull request, first Python script, first production deployment.
Reflection
AI changed what's possible. I went from zero GitHub experience to deployed production code in 4 hours.
Technology has come a long way. Now it needs to reach everyone.