lab

SpendSort

Confidence, gate, learn. The simplest honest agent loop, applied to expense categorisation

A month of card transactions categorised, with anything under the confidence gate held for review. An override is written to vendor memory, so the next run covers more from memory and calls the model less.

What it does

It categorises card transactions, tells you how sure it is, and only acts alone when it has earned the right to. Expense categorisation is the highest-volume, lowest-risk finance AI workflow, the right first agent to build.

Four nodes: normalise the vendor, check memory, categorise if it is a miss, and route on confidence. The memory is what makes the pattern worth shipping. A confirmed categorisation becomes a hit next month, so the gated fraction falls as the system learns rather than staying constant forever.

Repo

Stack

Python
LangGraph
FastAPI
SQLite
React

All data in this project is synthetic.