Aurelio is an open-source wealth-management app with an AI advisor

Personal-finance apps that connect to brokers and give AI-written advice are a paid category, with Monarch, Rocket Money and Copilot charging monthly fees for a dashboard the user cannot run on their own computer. Aurelio is a new MIT-licensed project that does the same work locally, and it has reached 701 stars on GitHub in three days.

The project is at LosaLosSantos/aurelio-finance. It is a wealth-management and financial-planning app that runs on the user's own machine and keeps every record in a local SQLite file at backend/data.db. The AI side is optional, and when it is on, the calls go through OpenRouter with the user's own key.

What the app covers

The project page lists seven areas it handles: wealth across banks and brokers, a portfolio priced from the market with cost, gain and dividends, a look-through view of what each fund holds, recurring investments and real assets with their financing, cash flow and the return each goal needs, a chat called Ask Aurelio that answers from the user's records and proposes every change as a card the user confirms, and an analysis mode where an analyst and a confidant argue over the portfolio and a synthesis says what to look at first.

The app runs on macOS, Linux and Windows. A --demo flag starts it on an invented household so a reader can look at it before giving it any real numbers. The code is Django on the backend and a Node 22 frontend, with uv run pytest and npm test as the suites.

What the AI calls cost

The repository lists its own measured prices with Claude Opus 5.5 as the default model on OpenRouter at $4 per million input and $20 per million output tokens:

• A first question, answered in one round: $0.11.
• A follow-up in the same session: $0.017.
• A request for advice with one web search: $0.16.
• A full analysis: $0.33 to $0.51.

The project repeats that it does not include any AI subscription of its own and that the OpenRouter key stays on the user's machine. A reader can swap in another model from the chat picker, which changes the per-call cost. The figures above are the author's own and will move with the OpenRouter price list.

For a reader paying $10 or $15 a month for a dashboard that sends their account data to a vendor, the trade is a local install, an OpenRouter key with a small balance, and a project that is still three days old. Backups of the local file sit next to it when the schema changes. The app's own disclaimers are stated plainly in the README, including that the AI advisor is optional, that an analysis costs more than a question, and that the user confirms every change the chat proposes.

Source

The repository is at LosaLosSantos/aurelio-finance, under an MIT licence.