Building Tattva AI: 24 Hours to a Functional AI Agent
Published on
Reading Time
2 mins
The Mission
At the COMSCI '26 hackathon hosted by SRMIST Ramapuram, my team and I tackled the digital epidemic of fake news. We built Tattva AI, a WhatsApp-based assistant that fact-checks messages, images, and voice notes in real-time.
The Technical Build
The system uses NLP to extract core claims from multi-modal inputs. Once a claim is identified, the engine performs automated evidence retrieval to cross-reference the information against verified databases, returning a credibility score directly to the user's WhatsApp interface.
The Outcome
We are proud to have secured the Runner-Up position! Building a multi-modal verification tool in a 24-hour sprint pushed our limits in API integration and rapid prototyping.
Key Achievements
Multi-Modal Support: Successfully processed text, image, and audio triggers.
WhatsApp Integration: Created a frictionless UX for non-technical users.
Silver Medalist: Recognized for technical execution among 24-hour builds.
