Smart India Hackathon

Kautuhala

Turning Scientific Knowledge Into Public Discovery

An intelligent scientific knowledge and outreach platform. It brings an institution's scattered reports, datasets, photographs, videos, and expedition records into one connected, searchable space, and uses AI to turn research into stories the public can understand.

SIH PS 26063 FastAPI PostgreSQL pgvector AI / LLM Docker

View Memories

Smart India Hackathon

26063

Problem Statement ID. Kautuhala was built as an MVP for this Smart India Hackathon problem statement.

The Problem

Research institutes hold thousands of files spread across websites and storage systems. Finding the right information is hard, and turning valuable research into something the general public can follow takes a lot of manual work.

What Kautuhala Does

Makes knowledge easy to discover

Research, expeditions, datasets, publications, photos, videos, and activities are organized and connected, so users can see how they relate to one another.

Makes knowledge easy to communicate

AI turns an authoritative research report into a website article, social-media posts, and visual content. An administrator reviews it before publishing.

Search finds what is relevant. Relationships reveal what is connected.

System Lifecycle

IngestUpload, bulk upload, or URL
ProcessExtract content and metadata
ConnectLink related resources
DiscoverSearch and explore
CommunicateAI outreach, human review

Connected Knowledge

Publications Expeditions Datasets Researchers Reports Photographs Videos Activities Articles

Example path: Publication → Expedition → Dataset → Researcher → Related Research → Media.

Two Experiences

Public Portal

No account needed
Search by topic, publication, expedition, researcher, dataset, location, or activity
Jump from one result into related knowledge

Institutional Workspace

Secure admin login and dashboard
Upload, manage metadata, and relationships
Generate outreach, review, edit, and publish

AI Outreach Studio

Website Article

A reader-friendly article written from the original report.

Social Media Posts

Platform-specific text for each channel.

Visual Content

Graphics for social media.

AI works from the institution's stored source material; the LLM is never treated as the source of truth. Every output goes through Generate → Review → Edit → Approve → Publish.