Overview
Auden AI is a multi-tenant, AI-powered personal-finance platform for people who track their money where they already talk. Users log income and expenses conversationally through Telegram, WhatsApp, and Discord using plain language, and every entry syncs to a web dashboard for analytics, history, and exports. It is built for individuals who want the friction of bookkeeping to disappear into the chat apps they use daily. Currently in private beta.
The Problem
Manual expense tracking fails because opening a dedicated app in the moment is too much friction, so transactions go unrecorded and financial pictures drift out of date. Auden AI replaces the spreadsheet-and-receipt workflow with a message you can send the instant money changes hands.
Key Features
- Conversational logging — users record transactions in natural language such as "lunch was 65k" or "/out boba 14k" without opening a separate app or filling out a form.
- Automatic categorization — the system reads each message and assigns every transaction to a category on its own, so entries are organized as they arrive.
- Multi-channel input — the same account accepts entries from Telegram, WhatsApp, and Discord, letting users log spending from whichever messenger is already open.
- Web dashboard — every logged transaction syncs to a web dashboard where users review analytics, browse full history, and export their data.
- RAG financial insights — a Retrieval-Augmented Generation pipeline fuses conversation history, transaction patterns, and user preferences to return context-aware financial insights.
- End-to-end type safety — TypeScript types are generated directly from the Rust structs, keeping the API contract and the web client in sync across the stack.
How It's Built
The backend is a Rust service built on Axum with SQLx for database access, exposing the API that every messaging channel and the web dashboard consume. TypeScript types are auto-generated from the Rust structs via ts-rs, giving the web client end-to-end type safety without hand-maintained interface definitions. Context-aware insights are powered by a Retrieval-Augmented Generation layer that draws on conversation history, transaction patterns, and stored user preferences. The platform is multi-tenant, isolating each user's data behind a shared API.
Tech Stack