Back to work

Project details

Trading Research

Stock Boy

A fork of Vibe-Trading rebuilt around personal trading context, IBKR connectivity, daily alerts, and automated trading workflows.

Live project
Published
Role
Fork adaptation, trading workflow design, local CLI/API validation, and Python pipeline engineering
Stack
Python / FastAPI / React
Historical market street illustration with crowded stalls, storefronts, and people trading goods

Intent

Stock Boy starts from a fork of HKUDS/Vibe-Trading, then narrows the product around my own trading and investment situation. The goal is not a generic trading demo; it is a personal operating system for seeded equity research, turning a market question into repeatable evidence, watchable signals, and actions that can be audited.

System

One side handles deterministic research machinery: seed validation, source-health summaries, run manifests, harvested evidence, scored payloads, CSV outputs, provenance, and final report packs. The other exposes CLI, API, web, and MCP surfaces around market-data loaders, strategy generation, backtests, reports, agent teams, an IBKR connection, daily alerts, and automated trading capabilities.

Current direction

The public entry presents the operating system for research, not investment recommendations. The useful evidence is the adaptation loop: a forked trading workspace made hyperspecific, connected to brokerage infrastructure, and constrained by provider checks, report artifacts, alerts, and metadata that make each automated action inspectable.