← Back to Blog
Development

Building my own stock research dashboard

Building my own stock research dashboard

I built a small private dashboard called Signal Desk AI for the handful of stocks I follow and plan to purchase.

It doesn't generate buy or sell signals. It's a workspace — a watchlist of companies across a few markets and risk levels. Research is on-demand rather than continuous: price history and international market data come from EODHD, company facts, statements and news from Financial Datasets, both pulled only when requested, with usage counters so I don't burn through API credits.

Signal-Desk-AI

After I buy something through my broker, I record it — either by hand or by importing a Monex trade-history CSV. Signal Desk tracks multiple purchase lots per company, account type, currency, execution price, fees, and the running average cost. Every stock also has a notes field for things I want to remember.

Signal Desk doesn't place trades, doesn't store brokerage passwords, and doesn't present advice.

There's an official read-only MONEX MCP Server, and I'm working on an integration where Claude or ChatGPT acts as an intermediary: the AI reads my actual holdings from Monex, stages a snapshot inside Signal Desk, and shows me what would be added, changed or removed. Nothing is applied until I confirm it.

Everything up to this point works: the research dashboard, the purchase ledger, the CSV import, the notes, the staged sync screen waiting for a live feed.

I'll post a link to the public repo when it's ready.

Share:

Comments

No comments yet. Be the first to write one.

Commenting as a guest.Sign in