Team 24

Context to Action by Ipsita

Context to Action by Ipsita is a self-learning personal AI agent that turns conversations into decision intelligence. It remembers what you learned and why, connects past decisions with new insights, spots contradictions, and learns from feedback to recommend meaningful actions with your approval.

Check your own project

Community votes recognize the build and do not affect official judging.

Review the project

Start with the source code, then open the demo or video if available.

View GitHub repository
Project description
Context to Action by Ipsita - Your Personal Decision Intelligence AgentIt is a self-learning personal AI agent that turns everyday conversations into decision intelligence. The Problem: We attend meetings, workshops, and networking events, make important decisions, and commit to follow-ups. But over time, we forget what we learned, why decisions were made, and what actions we promised to take. Traditional meeting assistants summarize conversations but don't always connect past decisions with new knowledge. The Solution: I built an AI agent that captures live conversations or analyzes uploaded transcripts, extracts decisions and their reasoning, remembers context across sessions, detects contradictions, and recommends meaningful actions. Users can ask questions with source-backed answers, fact-check information, and review, edit, and approve follow-up emails before sending. The agent improves its recommendations using feedback and reviewed preferences. Technology Stack: Electron: Native Mac desktop app with live listening, alongside a separate web app. Cognee: Connects decisions and reasoning through a knowledge graph. Firedrill: Tests agent safety through simulated approval and failure scenarios. SingleStore: Stores conversations, insights, and decisions in the cloud. Superlinked: Ranks historical evidence for more relevant retrieval. Differentiation: Beyond summaries, it remembers why decisions were made and spots contradictions.
Tools used