Fortuna

Screenshots

Problem Statement

There is a reason that brokers and sales-led financial models have higher success and satisfaction rates. Natural language is the interface of humans and thus commerce.Despite other attempts at simple use cases of "talk to a chat bot and have it do a thing in code", they failed to really drive at the heart of how retail financial modelling is done, especially when it comes to digital assets.Retail has rarely ever cared about the underlying financial engineering they benefit from unless it goes wrong. The idea that this should change to bring digital assets and its benefits to the masses is faulty. Access to concept of yield should not require an understanding of staking, yield farming, its overt risks, or complex financial mechanisms.Fortuna is a natural language chat interface specifically designed to remove any kind of notion of underlying understanding. We do this by modelling 2 core concepts that most retail investors are driven by:their immediate financial access needs, such as to send someone or something monetary value (not an address or computer)their long term financial modelling needs such as ensuring they keep most of their savings in an account or financial institution that fits with their trust, UX, and financial needs.Fortuna does this by abstracting all of these concepts into natural language, and providing AI agents with the tools they need to execute these models

Solution

We used account abstraction with Pimlico and Safe, to be able to execute transactions on behalf of a user. We used FastAPI for the backend and Next.js for the frontend. We used Dynamic for wallet connection.SAFE modules allow us to structure new ideas such as delegate AI agents