Skip to content
ENES
← Projects

finport · python · open source

Python libraries

finport: open-source Python libraries for investment work. Portfolio optimization, a point-in-time data store and a SOFR rates engine, built on one rule: every result shows what it rests on.

portfolio · cvxpy · mcp·Published

Optimization Engine

Multi-asset portfolio optimization engine — mean-variance, risk parity, HRP, Black-Litterman, CVaR — packaged on PyPI, with a Streamlit UI, CLI, and MCP server.
Read case study↗
Hover for more
macro data · SEC · point-in-time·Building

Data Pipeline

A local store of public economic and company data that keeps every revision, so a backtest can read a series as it was known on a past date. 1,293 curated macro series across 44 economies, plus SEC filings and XBRL facts.
View repo↗
Hover for more
rates · SOFR · mcp·Building

Rates Engine

Deterministic SOFR rates engine: curves, futures with convexity adjustment, swaps, hedging and options, plus USD/MXN. Every result returns the number and the evidence it rests on, with no AI at runtime.
View repo↗
Hover for more

Overview

finport is the family name for the Python libraries I build for investment research and publish in the open. They share conventions: deterministic results, no AI at runtime, typed APIs, and an MCP server wherever an agent needs to call the library as a tool.

They also share one rule. A result is only useful if you can check what it rests on. The optimizer returns the solver and the constraint compliance with every allocation. The rates engine returns the evidence behind each price. The data store keeps every revision, so a backtest reads a series as it was known on the day.

Status

One library is on PyPI today: finport-optengine. The data pipeline and the rates engine are in active development, and their repositories are public.