NeuroTrade, a local-first AI trading operator for Solana powered by Pyth Pro market data

Team: @SCP , Submitted: March 31, 2026


NeuroTrade is a local-first AI trading agent that uses real-time Pyth Pro price feeds for BTC, ETH, and SOL. It streams prices, computes indicators, and lets an AI model generate BUY/SELL/HOLD decisions, which can be simulated or executed via Jupiter api, while keeping all keys and execution on the user’s machine.


What It Does

NeuroTrade is an experimental trading system that runs locally and explores AI-driven decision-making using pyth pro data. It focuses on making the full loop from fetching data, reasoning, and onchain execution in a transparent and controllable environment.


Pyth Features Used

:check_mark: Price Pro


Links

Live Demo: Runs locally
Source Code: https://github.com/SCPassion/neuro-trade
Video Walkthrough: https://drive.google.com/file/d/13FpiyfRik15hzBiMAy3j4KkIlYCbQGDi/view


Screenshots / Media

Static captures of the three main pages (same assets ship in the repo under docs/screenshots/).

Dashboard

NeuroTrade dashboard — live chart, decision card, risk panel, and engine controls (paper mode)

Strategies

NeuroTrade Strategies — active profile, signals, guardrails, and limitations

Settings

NeuroTrade Settings — setup readiness, local secrets, and trading parameters


Tech Stack

Frontend:
React 19, TypeScript, Vite, Recharts

Backend:
Express 5, WebSockets (ws)

Blockchain / Trading:
Solana (@solana/kit), Jupiter execution API

Data:
Pyth Pro WebSocket (real-time price feeds)

AI / Reasoning:
OpenAI API

Storage & Validation:
better-sqlite3, Zod


Content Contributions (Required)

Not fully completed — sharing this project mainly for experimentation and learning during the hackathon. :grin:


Licensing

This project is licensed under Apache 2.0


Eligibility Confirmation

:check_mark: I am 18+ years old
:check_mark: I am not located in an OFAC-sanctioned jurisdiction
:check_mark: I confirm this is an original work created during the hackathon period
:check_mark: I have read and agree to the Terms & Conditions

Ps. This project is designed for fun. Do your own research before any trading.
Not financial advice.