What we know, exactly.

Three categories, nothing hidden:

Account — if you sign in with Google: your name, email address, and avatar, plus the session records needed to keep you signed in. Paper-trading state — your simulated account (balance, positions, orders, history, journal notes), watchlists, and saved setups, synced so your other devices see them. Product events — things like "page viewed," "order submitted," "bot run," with a random device identifier, so we can understand which features matter and improve them.

We take no payments today, so we hold no payment data. We never ask for and never store card numbers, bank details, tax or government identifiers, or credentials to any real brokerage. Signed out, your paper account lives only in your browser's localStorage and never reaches our servers.

Server-side data is stored in MongoDB Atlas. Client-side, we use localStorage for your practice account, preferences (like data mode), and the anonymous device id — plus the cookies Google sign-in requires for your session. Market quotes are proxied through our servers from Yahoo Finance; your name and email are not part of those requests.

We use your data to operate the features you use (sync, sign-in), to debug problems, and — via aggregate product events — to decide what to build next. We do not sell or rent personal data, and we run no third-party advertising trackers. The only parties that touch data are the infrastructure providers that host it (e.g. our database and hosting), Google, which handles the sign-in itself and — on Chrome and Edge — the speech recognition behind the free voice engine, and — if you use the voice co-pilot or an AI explainer — the AI providers named in clause 05.

The co-pilot has no brain of its own. Whichever engine you run, it is shown the same picture of your workspace — symbol, indicators, alerts and their notes, watchlist, paper orders — plus the computed read of the live chart. The free engine attaches both to every turn; the premium engine is handed them whenever it reads your screen, which it does from the opening greeting onward. Your name, email, and session are not part of any of it — but everything you say to the co-pilot is.

On the free voice engine, your speech is turned into text by your browser's built-in recognition: on Chrome and Edge — the browsers this engine actually works in — that means the raw microphone audio is uploaded to Google's speech service. The transcript then goes with the context above to OpenRouter, which routes it to a free community model. Free models are only served to accounts that allow prompt logging, so assume the provider stores those prompts and may use them to train future models.

The premium voice engine streams your microphone audio from your browser straight to OpenAI, and sends the same workspace and chart context down the same connection — so it discloses more than the free engine, not less, just to one named company instead of two. The options explainers send the trade's parameters (strategy, strikes, break-evens — not who you are) to OpenAI too. If you would rather none of this leaves your machine, don't start the co-pilot and don't press explain; nothing else on lazybull talks to a model. Moving the free engine onto a paid, no-retention tier is the fix we want, and this clause changes the day it lands.

Account data is kept while your account is active. To delete your account and everything attached to it, email hello@lazybull.us from your sign-in address and we will remove it. Signed-out practice data can be cleared any time via your browser's site-data controls (or the in-app reset button).

If this policy changes materially, the effective date above changes with it. Questions, requests, or complaints: hello@lazybull.us. The service terms live in the terms ledger.