##Inspiration## Inspiration
In India, justice is often a "legal mirage." Millions struggle to access their rights due to complex procedures, high costs, and lack of awareness. Everyday issues—like a landlord dispute or fraudulent online purchase—become overwhelming. We built Nyay-Sahayak to bridge this gap and provide citizens with “legal first-aid,” especially for marginalized communities.

What it does

Nyay-Sahayak is an AI-powered agent that helps citizens take the first step in navigating the legal system. Through a simple text or voice interface, users can:

  • Get Guidance: Step-by-step explanation of rights and procedures.
  • Generate Documents: Draft complaints, notices, or FIR checklists.
  • Navigate the System: Learn which authority to approach and what to expect.
  • Find Help: Access referrals to legal aid clinics, NGOs, and survivor support services.

Our focus areas include consumer grievances, landlord-tenant disputes, domestic violence remedies, and police encounters.

How we built it

  • Core orchestration with IBM Agent Development Kit (ADK).
  • AI “brain”: IBM Granite model, fine-tuned on simplified Indian legal codes.
  • Custom tools: ComplaintDrafter, JurisdictionFinder, RightsExplainer.
  • Built with ethics-first principles: privacy defaults, disclaimers, and escalation to human experts.

Challenges

  • Preventing misinformation in high-stakes contexts → mitigated with traceable, source-backed guidance.
  • Bridging the digital divide → long-term plan for multi-lingual voice and IVR access.

Accomplishments

  • Grounded design in socio-legal research.
  • Created an inclusive, ethical, citizen-centered tool.

What we learned

For AI to empower justice, it must be trustworthy, transparent, and co-developed with communities and legal aid providers.

What’s next

Pilot with NGOs and legal aid clinics, refine usability, and scale into a trusted “legal first-aid” layer for India.

Built with

IBM ADK, IBM Granite Models, Python, Kaggle, GitHub

Built With

Share this project:

Updates