Cryptelligence

Inspiration

The cryptocurrency sphere is steadily growing, and a lot of bad actors use it to try to avoid scrutiny.

What it does

Cryptelligence is an AI-assisted cryptocurrency investigation platform designed to help analysts turn a wallet address into a structured investigation.

An investigator can enter a wallet and an investigative question. Cryptelligence collects blockchain activity, maps relevant counterparties and transactions, and organizes the results into evidence that can support or contradict different explanations of what is happening.

Rather than only showing a transaction graph, Cryptelligence is designed to help answer questions such as:

What activity is this wallet involved in? Which counterparties are most relevant? What does the available evidence actually support? What is still unknown? What should the investigator look at next?

Our current prototype includes live Solana investigation support, transaction and entity visualization, evidence organization, and an early AI reasoning layer that helps synthesize findings into competing explanations and investigative next steps.

The long-term goal is to make crypto investigation less about manually jumping between explorers and more about maintaining an evolving, evidence-backed case.

How we built it

Cryptelligence uses a modular investigation architecture.

The frontend is built with React, TypeScript, Tailwind, and Cytoscape.js for the investigation graph.

The backend uses Python and FastAPI, with normalized models for: wallets and entities blockchain transactions evidence claims competing explanations investigative actions

For blockchain intelligence, we built a Solana investigation adapter that retrieves wallet activity and converts raw chain data into structured evidence that can be used by the rest of the system.

The AI layer uses Gemini to help interpret the investigator's objective, synthesize collected evidence, identify uncertainties, and suggest investigative directions. We deliberately separate deterministic blockchain facts from AI-generated inferences so the model is not treated as the source of truth.

The overall architecture is designed so additional blockchain, OSINT, monitoring, and specialized agent tools can plug into the same investigation workflow.

Built With

• and-cytoscape.js-for-the-investigation-graph.-the-backend-uses-python-and-fastapi
• css
• python
• tailwind
• typescript