# Team Pramaan

- **Event:** [Meta Global AI Developer Hackathon](https://cerebralvalley.ai/e/meta-ai-developer-hackathon)
- **When:** Oct 26, 12:00 PM – Nov 4, 9:00 PM (EST)
- **Where:** Online
- **Team:** [Ekta Sikka](https://cerebralvalley.ai/u/EKTAS)
- **GitHub:** https://github.com/ektasikka1989-cmyk/Pramaan
- **Demo video:** https://youtube.com/shorts/Su5WsNUSMm4
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2906

PRAMAAN — An Offline AI Co-worker for India's Frontline Health Workers

INTRODUCTION

Pramaan is an offline-first, AI-powered Android assistant designed to help India's approximately ten lakh ASHA workers deliver more organised, informed and accountable community healthcare.

PROBLEM

An ASHA may serve around 1,000 people while managing pregnancies, newborn care, immunisations, medicines, referrals and follow-ups.

Her biggest challenge is not just paperwork. She must consult hundreds of pages of government guidelines and SOPs, identify the correct instructions, verify her actions, maintain multiple registers and coordinate with ANMs and Primary Health Centres (PHCs).

With unreliable internet, limited time and numerous families depending on her, important follow-ups can be missed.

OBJECTIVE

Empower every ASHA with a reliable AI co-worker that simplifies fieldwork, makes official guidance accessible, improves recordkeeping and supports timely follow-ups—even without internet connectivity.

SOLUTION — PRAMAAN

Pramaan connects guidance, fieldwork and documentation in one application.

Its signature feature, Ask SOP, allows workers to ask questions in Hindi, English or Hinglish using a knowledge base of 68 official healthcare documents.

Unlike a conventional chatbot, answers include the exact document, page number and highlighted supporting passage, with access to the original PDF. This enables workers to inspect the evidence behind AI responses, even offline.

Pramaan also:

Converts voice notes into structured, worker-reviewed visit forms.

Maintains MCP cards, HBNC/HBYC records and registers.

Tracks immunisations, overdue visits, medicines and follow-ups.

Flags reported danger signs using predefined rules.

Prepares referrals and information for ANMs and PHCs.

Explains relevant guidance to families in accessible language.

Organises records and pending actions to support accountability.

POWERED BY META AI

Meta Llama 3.2 3B powers on-device language intelligence without continuous internet connectivity.

When configured and connected, Meta Muse Spark, Muse Voice Transcribe and SAM 3.1 extend online language, speech and vision capabilities.

Meta Muse Code supported development, integration, testing and optimisation.

IMPACT AND SCALABILITY

Pramaan has the potential to assist ten lakh ASHA workers and indirectly benefit millions of mothers, children and families.

Its architecture could extend to ANMs and other healthcare professionals through role-specific knowledge bases and workflows.

OUR VISION

No frontline health worker should be left without relevant guidance, organised records or actionable follow-ups simply because the internet is unavailable.

Pramaan supports trained healthcare workers. It is a hackathon prototype, not a diagnostic tool

## More from Meta Global AI Developer Hackathon

- [vrajhna](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2903)
- [Front-Row](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2904)
- [PartFidelity](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2905)
- [Iron & Pine Outdoor Co.](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2907)
- [Damian](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2908)
- [Sarvan](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2909)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2906. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt