Vibe Math Mcp
@apetta
A high-performance Model Context Protocol (MCP) server for math-ing whilst vibing with LLMs. Built with Polars, Pandas, NumPy, SciPy, and SymPy for optimal calculation speed and comprehensive mathematical capabilities from basic arithmetic to advanced calculus and linear algebra.
Overview
What is Vibe Math Mcp?
Vibe Math Mcp is a high-performance Model Context Protocol (MCP) server for mathematical calculations with LLMs. It provides 21 math tools across basic calculations, arrays, statistics, finance, linear algebra, and calculus, plus batch orchestration, using Polars, Pandas, NumPy, SciPy, and SymPy.
How to use Vibe Math Mcp?
Install with uvx vibe-math-mcp. Configure in Claude Desktop via Settings > Developer > Edit Config, or using claude mcp add with the stdio transport. Supports both published package and local development modes.
Key features of Vibe Math Mcp
- 21 mathematical tools across 6 domains
- Output control with 5 verbosity modes (full to final)
- Batch execution achieving 90-95% token reduction
- Built on Polars, Pandas, NumPy, SciPy, SymPy
- STDIO and HTTP transport options
- MCP-native with detailed parameter descriptions
Use cases of Vibe Math Mcp
- Evaluate complex expressions like "15% of 250"
- Compute matrix determinant of [[1,2],[3,4]]
- Integrate x^2 from 0 to 1 symbolically or numerically
- Calculate compound interest with varying frequency
- Chain multiple calculations for financial models or bond pricing
FAQ from Vibe Math Mcp
What dependencies are required?
Python, uv, and the libraries Polars, Pandas, NumPy, SciPy, and SymPy. Install dependencies with uv sync.
How do I install Vibe Math Mcp?
Run uvx vibe-math-mcp for the published package, or use uv run vibe-math-mcp from a local checkout for development.
What transport modes are supported?
STDIO (default for Claude Desktop and IDEs) and HTTP (for container testing via uv run python -m vibe_math_mcp.http_server).
Can I control the output verbosity?
Yes, every tool accepts an output_mode parameter with modes: full, compact, minimal, value, and final, saving up to 95% tokens in batch mode.
What is batch execution?
batch_execute chains multiple calculations in a single request, referencing prior outputs with $operation_id.result syntax. It handles dependency resolution and parallel execution automatically.
