Jacobian

•
375
•

Provides AI agents with composable mathematical operations and independently verifiable evidence for conjectures, computations, and formal proofs.

Jacobian empowers AI agents with a composable toolbox of mathematical capabilities, moving beyond opaque solvers to offer fine-grained operations for conjectures, counterexamples, exact computation, and formal proof. It ensures every step is visible through typed results and inspectable artifacts, emphasizing a deliberate trust boundary where only an operator-authorized checker can emit a 'VERIFIED' record. This design allows agents to construct objects, compute invariants, search for witnesses, and submit exact evidence, making the process of mathematical reasoning transparent and independently verifiable.

Key Features

01Independent verification of claims with inspectable evidence and artifacts
02Broad capabilities across polynomial algebra, graph theory, SAT/SMT, and Lean formalization
03Composable mathematical operations tailored for AI agents
04Support for exact computation, formal proof, and counterexample generation
053 GitHub stars
06Interfaces as an MCP server, CLI, and Python library

Use Cases

01Provide a trusted environment for exact computations and formal proofs in agent-driven workflows
02Enable AI agents to perform complex, verifiable mathematical reasoning
03Automate the discovery and formal verification of mathematical conjectures and counterexamples