Adhikaar – AI Welfare Copilot for Every Citizen

An AI welfare copilot that lets citizens discover schemes, extract data from documents via OCR, and auto-apply using voice in their native language, simplifying access to 4,600+ government benefits.

Comment

Inspiration

What it does

How we b## Inspiration

Millions of Indian citizens are eligible for government welfare schemes but never receive benefits due to lack of awareness, language barriers, complex eligibility rules, and difficult application processes. Existing portals provide information, but they do not guide users through the full application journey.

We built Adhikaar to act as an AI welfare copilot that simplifies discovery, eligibility checking, document understanding, and application submission through a single voice-driven interface.

What it does

Adhikaar is a voice-first AI copilot that helps citizens discover eligible government schemes and apply instantly in their native language.

The system:

  • identifies relevant schemes from 4,600+ government programs
  • extracts details from Aadhaar, PAN, and other documents using OCR
  • checks eligibility automatically
  • auto-fills application forms using intelligent browser automation
  • supports multilingual interaction across major Indian languages
  • allows users to apply for themselves and family members
  • provides a dashboard to track applications

This makes welfare access faster, simpler, and more inclusive.

How we built it

The frontend is built using Next.js 16 with Tailwind CSS and modern UI components for a responsive experience.

We integrated:

  • Groq Llama 3.3 70B for reasoning and scheme assistance
  • Google Gemini Vision for OCR-based document parsing
  • Sarvam AI for multilingual speech recognition and synthesis
  • LiveKit Cloud Agents for real-time voice interaction
  • myScheme.gov.in API for government scheme discovery
  • Playwright for automated application form filling

Together, these components create an end-to-end AI-powered welfare application pipeline.

Challenges we ran into

One major challenge was handling multilingual voice interaction reliably across different accents and languages.

Another challenge was designing a universal automation workflow capable of filling different government forms dynamically.

We also worked on improving document parsing accuracy so extracted data could be reused across multiple applications without repeated manual entry.

Impact

Adhikaar reduces the gap between citizens and government welfare delivery by transforming complex scheme access into a simple voice conversation.

It enables:

  • faster scheme discovery
  • automated application workflows
  • inclusive access for non-English speakers
  • support for rural and low-digital-literacy users

The platform has the potential to scale nationwide and act as a digital welfare assistant for every citizen in India.uilt it ho

Built With

  • framer-motion
  • google-gemini-2.0-flash-(vision-+-realtime)
  • groq-llama-3.3-70b
  • gsap
  • livekit-cloud-agents
  • myscheme.gov.in-api
  • next.js-16
  • playwright-automation
  • python-voice-agent
  • react-19
  • sarvam-ai-(stt/tts)
  • shadcn/ui
  • tailwind-css
  • vercel
  • vercel-ai-sdk
Share this project:

Updates