ChainLens
ChainLens is an AI workspace transforming natural language queries into interactive research views with real-time charts and risk data.
YouTube Video
Project Description
ChainLens is an AI-powered crypto intelligence workspace that turns crypto research questions into structured, visual analysis.
Users can ask conversational prompts like “Give me a quick brief about Cardano today,” “Compare ADA, BTC, and ETH,” “What is ADA?”, “Show risk signals for SNEK,” or “What happened this week in the Cardano ecosystem?” Instead of returning only a text answer, the agent updates a live intelligence canvas with purpose-built views such as summary cards, comparison tables, risk panels, signal lists, timelines, project watchlists, and token research blocks.
The project focuses on controlled Generative UI. The agent decides which backend tool and view best match the user’s question, but the frontend only renders approved crypto UI components. This keeps the experience predictable, safe, and demo-ready while still feeling dynamic and conversational.
For explicit interactive UI prompts, ChainLens uses a deterministic crypto risk rendering path so the demo does not depend on the model hand-writing fragile UI JSON. Normal research flows update shared agent state directly, which lets the canvas render reliably even when the assistant response is brief.
The demo uses mock crypto data for assets, market snapshots, Cardano ecosystem projects, recent events, token documents, and risk signals. It is designed to demonstrate the interaction model and should not be treated as financial advice.
https://github.com/GinoLlerena/Generative-UI-Global-Hackathon-Starter-Kit
Prior Work
The project builds on our team’s prior experience working with AI agents, frontend tooling, and structured UI workflows. Before the hackathon, we were already familiar with concepts like agent orchestration, tool calling, constrained generative UI, and using predefined UI components instead of allowing completely open-ended UI generation.
For the hackathon, we used the provided starter template as the base UI/application structure and adapted it to our project needs. The main work created during the hackathon includes the project concept, demo flow, agent behavior, UI block contract, frontend tool integration, and the specific implementation for the Crypto Intelligence Copilot demo.
No proprietary company code, private internal systems, or confidential business logic were included in the hackathon submission. Any prior knowledge used was general technical experience, not private source code or restricted designs.