A portfolio tracker that runs on your machine#
Pyx is a native desktop app for watching your investment portfolio. It pulls holdings and transactions from your brokerage, fetches prices from Yahoo Finance, and renders everything as fast charts and tables. All of it lands in one encrypted SQLite database on your own computer.
That last part is the point. Most portfolio trackers are websites. You hand them your accounts, they hold your data, and you trust their servers, their employees, and their business model. Pyx inverts that. Your data stays with you. The app talks to a short, auditable list of hosts and nothing else.
Pyx is a research workspace, not a trading terminal. It does five things:
- Syncs holdings and transaction history from your brokerage through SnapTrade.
- Fetches live and historical prices for what you own.
- Computes returns, allocation, and cost basis the way an institution would.
- Holds your research: charts, tags, and markdown notes linked to your positions.
- Keeps every byte of it on your machine.
It does not place trades, give advice, or file your taxes.
What makes Pyx different#
Local first. Your portfolio lives in an encrypted SQLite file. Brokerage secrets live in an encrypted secrets file sealed by a key in the operating system keychain, not in a config file. Network egress is allowlisted in source code, so the app cannot talk to hosts that are not on the list. There is no telemetry to turn off because none exists.
Fast. Pyx is written in Rust and draws its charts on the GPU. Twenty years of daily price history pans and zooms without stutter. Tables stay smooth with full multi-year transaction histories in them.
Keyboard driven. CmdK opens a command palette that reaches every view, action, and ticker. A Vim keymap ships in the box. Most of the app is reachable without touching the mouse. See the command palette.
Text you own. Configuration is layered TOML files with hot reload. Chart workspaces save as TOML files you can version, diff, and share. If you keep a dotfiles repo, Pyx will feel familiar. See configuration.
Organized your way. Tags group securities you own or just watch, each with fundamentals, upcoming earnings, and a compare chart against your benchmark. Saved views remember any filter and sort combination on the holdings table. The app bends to your workflow, not the reverse.
A research journal built in. Notes are markdown with live preview. They link to securities, accounts, and tags, and they pin to your charts on the day you wrote them, so the thesis and the outcome share a timeline. Attachments stay on your disk.
Works with your AI, on your terms. Pyx ships an MCP server, so Claude Code, Claude Desktop, Codex, or any MCP client can read your portfolio and run research over a local socket. Access is off by default, scoped by capability switches, and portfolio data is read-only to agents down to the database layer. No bundled chatbot, no portfolio data sent to an AI company by Pyx.
Who Pyx is for#
Pyx fits self-directed investors who already know what they own. The typical setup is index funds plus a handful of individual positions spread across a brokerage account and an IRA or two. You check your portfolio deliberately, not compulsively. You have read a privacy policy and closed the tab.
If you like tools such as Zed, Obsidian, or a tiling window manager, the design decisions here will make sense to you.
What Pyx is not#
Honesty saves everyone time. Pyx currently has:
- No trading. It is read-only by design; execution stays at your broker.
- No market-wide screener or backtesting engine. Researching and watching candidates happens through tags, notes, and the chart workspace.
- No options or crypto analytics. Options positions from a linked broker import as positions.
- One base currency at a time. Every portfolio total converts into it at that day's FX rate; native prices and transaction amounts keep their own currency.
- macOS (Apple Silicon), Windows, and Linux, all from the same download page.
If one of those is essential to you right now, the comparison with other tools is a frank look at what fits where, including where alternatives genuinely win.
How Pyx makes money#
Downloading and using Pyx with your own data is free. A subscription pays for brokerage sync, verified in-app updates, and agent access: $7 per month or $70 per year, one plan. Checkout runs through Polar, and a license key arrives by email. There is no account, no password, and nothing here is funded by selling your data. Details on pricing.