Back to feed
Dev.to
Dev.to
7/24/2026
Building a Telegram Nutrition-Scoring Bot for Swiggy Instamart: Architecture and Challenges

Building a Telegram Nutrition-Scoring Bot for Swiggy Instamart: Architecture and Challenges

Original: I Built a Nutrition-Scoring Bot Inside Telegram, and Swiggy Just Approved It

Short summary

A solo developer built a Telegram bot that scores nutrition for Indian packaged snacks ordered via Swiggy Instamart, assembling data from external food databases and LLMs when structured data is unavailable. The project was accepted into Swiggy's Builders Club. Key engineering challenges include caching untrusted LLM output, localizing the Nutri-Score formula for Indian products, and cutting reply time from 30 to 5 seconds.

  • •Telegram bot scores nutrition for Swiggy Instamart snacks using external databases and LLM fallbacks
  • •Accepted into Swiggy's Builders Club; built solo with Python, FastAPI, Redis
  • •Core challenges: caching untrusted LLM output, localizing Nutri-Score for Indian products, reducing latency from 30s to 5s

Generated with AI, which can make mistakes.

#ai-tools#ai-agents#industry-adoption
Read full article at Dev.to

Is this a good recommendation for you?

Comments

Failed to load comments. Please try again.

Explore more

I built a Telegram bot that counts calories from food photos. It confidently called soup "berry compote"
Dev.to
I Built an AI App. Eight Months Later, It Became a Skill
Dev.to
Building a Telegram AI Agent for Personal Use
Dev.to
Taming Open Food Facts: building a fast, searchable nutrition database
Dev.to