marco todaro·Ai engineer
projects/

Bussola

2022 – present
Self-hosted personal finance, spending and net worth together

Why I built it. My money sits in many places: bank accounts, cards, payment apps, a broker, an exchange, cash. Budgeting apps see the spending, portfolio trackers see the investments, and the real total ended up rebuilt by hand in a spreadsheet. I wanted:

  • control of my data: it is sensitive, so it stays on my own machine, and no service ever gets my bank credentials;
  • a clear, simple interface: record an expense in seconds and see where the money is at a glance, without double-entry bookkeeping;
  • as little manual entry as possible, because typing every movement is why people give up.

I started it in 2022 and used it for years with only the core features. In 2026 I finished it with Claude Code and released it as open source.

Main features

  • Every bank, card, payment app, broker, exchange and pocket of cash adds up to one total.
  • Income and expenses by month and category, next to the portfolio.
  • Investments split across four pillars: cash, emergency fund, planned expenses, long term.
  • Transactions from bank notifications: MacroDroid or Automate forward them, an AI model (local or any OpenAI-compatible API) drafts the transaction, you accept it in one tap. Bank CSV import too.
  • Automatic prices for ETFs, stocks and crypto, and a monthly snapshot of the portfolio.
  • Wallets shared with family, read-only or read-write.
  • Passkeys, an installable app with push notifications, six languages, and a bank mode that blurs every amount.

The portfolio split across the four pillars, with its total and its value over time

React · Express · PostgreSQL · Docker · LLM
site ↗demo ↗github ↗

wiki-assistant

Chat over a personal markdown wiki

An OpenWebUI ↔ OpenCode bridge for querying a personal markdown wiki with an LLM. Deliberately no RAG and no embeddings: the model reads the actual files, orienting itself from a root index.md (based on LLM Wiki idea).

OpenWebUI · OpenCode · Docker · LLM Wiki
github ↗

Payment Reconciliation

Bank transfers matched to policies · Wide Group

An AI system that matches incoming bank transfers against customer insurance policies, handling the messy free-text references humans put in payment descriptions. Reduced manual reconciliation time by roughly 80%.

Python · scikit-learn · Kafka · PostgreSQL

siyuan-mcp

MCP server for the SiYuan note app

An MCP server that exposes a SiYuan knowledge base to LLM clients such as Claude Code, down to individual notebooks, documents and blocks, so an agent can read and write personal notes as a first-class tool rather than through copy-paste.

Python · MCP · uv
github ↗

Scraper Agents

Agentic data extraction at scale · Wide Group

A scalable scraping architecture built on AI agents and a custom MCP server. Agents handle the parts that change: page structure, selectors, edge cases. The surrounding workflow stays deterministic, so runs are reproducible and failures are debuggable instead of mysterious.

Python · MCP · LangChain · Kafka · Docker

LLaVA / LLaVA-NeXT adapters

MSc thesis: multimodal adapters

Improved multimodal LLM performance using a C-abstractor adapter between the vision encoder and the language model, trading a small amount of spatial detail for a large reduction in visual token count.

PyTorch · Python · MLLM
LLaVA ↗LLaVA-NeXT ↗

NaviSync

Navidrome → phone music sync

Android app that mirrors an entire Navidrome library (or any Subsonic-compatible server) into a local folder any player can read. Playlists export to .m3u8 with relative paths, and starred tracks sync both ways.

Not rsync: Subsonic only speaks HTTP, so the sync is idempotent logic on top of the API, index the server, diff against the folder, and pull only what is missing or changed via the download endpoint, which returns the original file rather than a transcode.

Kotlin · Android · Subsonic API
github ↗

muxlib

Self-hosted music library manager

Downloads YouTube audio as MP3, writes the metadata, and files it into a personal library. Driven from a Telegram bot, so adding a track is one message from a phone: send a link and it is either delivered back as a file or saved straight to the server, depending on who is asking.

Python · yt-dlp · Telegram Bot API
github ↗

Homelab

Self-hosted everything, across two houses

Two machines, in two different houses, running about twenty services behind a reverse proxy and joined by a zero trust network rather than an open port. Photos, music, audiobooks, ebooks, notes and files all live there rather than in someone else’s cloud, and a dashboard tells me at a glance what is healthy and what is not.

Part of it is stuff I wrote: a personal wallet app with its own API and Postgres database, a Telegram bot that reports how much the solar panels are producing, and a few smaller things that exist only because it was quicker to build them than to look for them.

It is also where anything new gets tried first. If a tool survives a few weeks of me actually depending on it, it stays.

The homelab dashboard: infrastructure and application tiles, each with a health status

Docker · Linux · Nginx Proxy Manager · Twingate · PostgreSQL

Earlier projects

University and side work
  • Smart Library: distributed system for fully autonomous, delocalised book lending across automated totems. IoT hardware, the architecture around it, and the loan logic.
  • Best Client: Android app that surfaces a customer’s rating while the phone is ringing, so restaurateurs know who is calling. Published on the Play Store.
  • Colors Zone: real-time COVID-19 restriction zone for any Italian region, with the rules attached to each colour.
  • Actor Detector: automatic detection and recognition of the actors present in a movie scene.
  • ML Algorithms: classification, regression, clustering, dimensionality reduction and ensembles, plus model selection and evaluation.
Java · Android · JavaScript · Python
smart-library ↗best-client ↗colors-zone ↗ml-algorithms ↗
h/l ←→ panej/k move⏎ open/ searcht theme? help