A local-first portfolio management application that lets investors import broker data into a single private workspace, display the data in multiple personalised views and support the analysis through an AI-assisted chatbot.
How I built it
Used Claude Code as a virtual cross-functional development team, creating specialised skill workflows for architecture, UI development, testing, governance and project management. Integrated MCP tools including Figma, shadcn and Playwright to connect product design, implementation and E2E testing.
01 — Problem
⚠️
Portfolio data split across brokers
Portfolio data is fragmented across brokers, while existing broker interfaces offer limited personalisation and cross-portfolio analysis.
02 — Product
💡
A local-first, personal analytics workspace
A local-first portfolio management application that consolidates broker data into a personalised analytics workspace, complemented by an AI portfolio assistant.
03 — Key decisions
🔀
Local-first, starting with IBKR
Local-first architecture · IBKR-first MVP · Broker abstraction · AI grounded in portfolio data