MIT-licensed MCP server (also a CLI and Python API) that gives agents a typed, searchable vocabulary of bounded mathematical operations — exact results with approximation and uncertainty made explicit.
DSH integration
Compatible
Author-claimed
Safety audit
Unaudited
Last verified
2026-08-29
License
MIT
01What can it help you accomplish?
Give an AI agent a searchable, typed vocabulary of mathematical operations to discover and run
The discovered operation's signature (`math.find`) or the typed result of one bounded contract (`math.run`)
Agents and developers who need exact, composable math primitives instead of large domain solvers
Compose exact, explicit mathematical results into agent workflows without a proof strategy or solver
Results that are exact where claimed, with approximation, incompleteness, or uncertainty made explicit
Reasoning and theorem-proving agents and tool builders who want bounded, reusable math postconditions
02How to install into DeepSeek Harness
Not specified by the author
03DSH integration and capability boundaries
Exposes typed mathematical operations as an MCP server that agents (including DeepSeek Harness) connect to; the same library also ships as a CLI and a native Python API.
math.find — discover an operation
a natural-language or typed query for a mathematical operation→the matching typed operation's signature
math.run — execute one bounded contract
a discovered operation plus its typed arguments→the typed result of exactly one bounded mathematical contract
CLI and native Python API
the same mathematical library, called from a shell or Python→the same typed results, outside the MCP server
04Who is it for? When not to use it?
Good for
- Agents and developers who need exact, composable math primitives instead of large domain solvers
- Reasoning and theorem-proving agents and tool builders who want bounded, reusable math postconditions
05Compatibility, maintenance and safety notes
- Each call is scoped to one bounded, reusable mathematical postcondition; Jacobian does not prescribe a workflow or proof strategy.
- Results are exact where claimed; when a result is approximate, incomplete, or uncertain, Jacobian makes that explicit rather than hiding it.
MIT · actively maintained (latest release jacobian-v0.15.1, 2026-08-27)
06Frequently asked questions
How do I connect Jacobian to DeepSeek Harness?
Jacobian is an MCP server, so add it to your DeepSeek Harness MCP configuration; the same math library is also callable directly from the CLI or the native Python API if you prefer not to run the server.
What does `math.find` do versus `math.run`?
`math.find` discovers a typed operation from the searchable vocabulary, while `math.run` executes exactly one bounded mathematical contract and returns its typed result — one call, one postcondition.
Are the results exact?
Results are exact where claimed. Where a computation is approximate, incomplete, or uncertain, Jacobian makes that approximation, incompleteness, or uncertainty explicit rather than hiding it.
Does Jacobian replace a proof assistant or SMT solver?
No. Each operation establishes one stable, reusable mathematical postcondition rather than prescribing a workflow or proof strategy, and it does not expose large domain solvers — it is a composable vocabulary of bounded operations.
What languages or runtimes are supported?
The mathematical library is available three ways: the MCP server for agents, a CLI, and a native Python API, so you can call it from DeepSeek Harness or any Python toolchain.
07Related DSH workflows
ruflo
by ruvnet
🌊 The original agent harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, federation, vector RAG integration, and native Claude Code / Codex / Hermes and many more Integrated
reactive-resume
by reactive-resume
A one-of-a-kind resume builder that keeps your privacy in mind. Completely secure, customizable, portable, open-source and free forever. Try it out today!
everos
by evermind-ai
One portable memory layer for every AI agent: local-first, Markdown-native, user-owned, and self-evolving across apps, tools, and workflows.
openpencil
by zseven-w
The world's first open-source AI-native vector design tool and the first to feature concurrent Agent Teams. Design-as-Code. Turn prompts into UI directly on the live canvas. A modern alternative to Pencil.
08Data and sources
Jacobian is an MCP server that gives AI agents a searchable vocabulary of typed
An executable mathematical vocabulary for agents: discover one typed operation, run it, and compose its result.
This page is generated from the project’s public documentation, repository metadata and a structured parse of DSH Plugins; last verified on 2026-08-29. Found an error? Submit a correction.
Best DeepSeek Harness Plugins
Twelve plugins worth installing first — picked from the whole catalog, across every category.