Home / Projects / Scheme Finder
AGENTIC AI · NATIONAL

Every scheme you qualify for, in 4 questions.

Real data — 170 schemes

Problem Statement

Research question: Given a person's basic profile, can a system identify which government welfare schemes they likely qualify for, in a handful of questions rather than a manual search?

India has hundreds of welfare schemes, and most eligible citizens never discover the ones that apply to them. This project builds a conversational eligibility matcher against 170 real schemes scraped from myscheme.gov.in.

170
Real Schemes
85%
Tag Coverage
4
Questions Asked

Try it live

Runs on Hugging Face Spaces
Waking up the model — this can take up to a minute on the first visit today. It'll be instant after that.

Tag-based matching, not parsed rules

Eligibility criteria are free text, not structured rules. This matches on each scheme's real tags instead — and always shows the scheme's actual eligibility text so users can verify themselves.

A real data-cleaning step

251 unique raw tags had many spelling variants of the same category (PwD, Person With Disability). Manually grouped into 15 canonical categories before matching.