HN
Today

Show HN: SymDerive – A functional, stateless symbolic math library

SymDerive is a new functional, stateless symbolic math library designed with 'Agent-Native' AI reliability in mind, aiming to prevent AI hallucinations by enforcing a strict input-transform-output pipeline. It also serves as a bridge for physicists transitioning from Mathematica to Python by offering familiar Wolfram-style syntax atop Python's scientific stack. This project appeals to HN's interest in practical AI tooling and scientific computing innovation.

6
Score
0
Comments
#21
Highest Rank
4h
on Front Page
First Seen
Feb 4, 6:00 PM
Last Seen
Feb 4, 9:00 PM
Rank Over Time
30212528

The Lowdown

SymDerive is a novel symbolic math library created by a physicist-turned-quant, focusing on making mathematical operations more reliable for AI agents while remaining practical for human researchers. The library aims to address common issues with AI-generated code and ease the adoption of Python's scientific ecosystem.

  • Agent Reliability: The primary design goal is to enable AI agents to write more reliable code by enforcing stateless, functional pipelines, preventing issues like hallucinating state changes or getting lost in procedural scripts. This promotes a clear "Input -> Transform -> Output" workflow.
  • Python Transition: For physicists accustomed to Mathematica, SymDerive offers a familiar syntax (e.g., CamelCase, Sin, Integrate) while operating on the standard Python scientific stack under the hood, smoothing the migration process.
  • Technical Implementation: It functions as a functional wrapper around established libraries such as SymPy, PySR, and CVXPY, providing a unified, pipe-based workflow for both humans and AI.
  • Key Features: The library boasts "Wolfram Syntax" for intuitive use, a modular design allowing optional installation of heavier components like Symbolic Regression, and includes specialized physics tools (e.g., abstract index notation for General Relativity, Kramers-Kronig).

This opinionated library is positioned as a valuable tool for anyone building AI agents for rigorous mathematical tasks or researchers seeking a functional, familiar interface for their scientific computing needs.