Financial Models as Code

Orcaset has released an experimental Python library for building financial models as code, using type annotations to catch malformed models at write time. The library allows humans and AI agents to build, modify, and audit large financial models with confidence through strong typing, flexible data integration, and transparent, deterministic calculations.

This library is in experimental status and the API is likely to change. Orcaset is a Python library for building financial models as code. Type annotations catch malformed models at write time, so humans and AI agents can build, modify, and audit large models with confidence. Build and update with confidence — Strong typing promotes correct models and surfaces relationship dependencies, so you can modify models without breaking them. Flexible integration — Connect to any data source and embed in existing workflows by leveraging Python's ecosystem. Transparent and deterministic — No black boxes. Materialize and inspect query-specific traces to audit every calculation. Efficient construction — Reusable components and terse definitions mean faster iteration and less context rot for agents. Install from GitHub with uv or pip : uv add git+https://github.com/orcaset/orcaset-py pip install git+https://github.com/orcaset/orcaset-py Requires Python 3.14 or later. Add the orcaset-py skill to your coding agent by pasting in the following command: Install the skill from https://github.com/Orcaset/orcaset-py/tree/main/skill/orcaset-py Orcaset models are built by defining and combining line item classes. Values are materialized by querying over dates, with caching and circular references handled by the evaluation context. The snippet below defines a simple model with revenue, expenses, and income, then queries quarterly values through 2026. python from datetime import date from typing import Iterable from dateutil.relativedelta import relativedelta from orcaset import Context, Formula, Period, Span, SpanSeries, Stmt, Total, fixed width table, span, split daily, sum spans, Assumptions start date = date 2025, 12, 31 initial revenue = 100.0 revenue growth rate = 0.05 Model class Revenue SpanSeries : agg = sum spans 0.0 def spans self - Iterable Span : value = initial revenue for period in Period.seq start date, relativedelta months=1, day=31 : yield Span period, Formula.pure value , split daily value = 1 + revenue growth rate / 12 Expenses = span.scale Revenue, -0.7, name="Expenses" Income = span.sum Revenue, Expenses , agg=sum spans 0.0 , name="Income" Build a statement view stmt = Stmt Total Income, Revenue, Expenses Resolve and print formatted values ctx = Context qtrly periods = Period.list start date, relativedelta months=3, day=31 , date 2026, 12, 31 print fixed width table stmt.values ctx, qtrly periods Start 2025-12-31 2026-03-31 2026-06-30 2026-09-30 End 2025-12-31 2026-03-31 2026-06-30 2026-09-30 2026-12-31 Revenue 301.25 305.03 308.86 312.74 Expenses -210.88 -213.52 -216.20 -218.92 ---------------------------------------------------------------------- Income 90.38 91.51 92.66 93.82 Note that Revenue is defined on a monthly basis but queried quarterly. Orcaset automatically aligns, interpolates, and aggregates over partial periods. See the quickstart example /Orcaset/orcaset-py/blob/main/examples/quickstart/README.md for a step-by-step guide covering cells, series, context, and structured output. Orcaset is licensed under the Server Side Public License. You can freely use it to build internal models for underwriting, valuation, risk, or other analysis. See LICENSE /Orcaset/orcaset-py/blob/main/LICENSE for details.