# Optimatika Optimisation Service

> Optimatika Optimisation Service is a freemium MCP server for developer tools. MCP server that solves LP, QP and mixed-integer optimisation models for AI assistants. It runs as a hosted remote server, so there is nothing to install locally.

- **Kind:** MCP server
- **Pricing:** Freemium
- **Platforms:** API, self-hosted
- **Website:** https://www.optimatika.se/optimisation-service/docs/
- **MCP Registry name:** se.optimatika/optimisation-service
- **Categories:** [developer tools](https://ai-powered.com/categories/developer-tools)
- **Listing updated:** 2026-11-18

## About

The MCP server gives an AI assistant a solve tool for linear, quadratic and mixed-integer optimisation models. The model is passed as text in LP, MPS or EBM format, and the assistant gets back the solution state, the objective value and each variable's value by name. A second tool translates models between LP, MPS and EBM, and a third reports how large a model may be and which solvers are available.

The server runs on Optimatika's shared instance at a Streamable HTTP address, and it is stateless, so each request stands alone. Long solves can report progress to a client that asks for it, and ending the call aborts the solve. The same service also exists as a container image that can be run on your own infrastructure, where models do not leave your environment. The shared instance is intended for testing rather than production, and the docs advise against sending it confidential models.

This suits people who already have an optimisation problem to model, such as planning or allocation, and want an assistant to formulate and solve it without installing a solver. Use without a licence key is free but limited to models of up to 1,000 variables or constraints and shorter solve times. A licence key from an Optimatika subscription raises those limits and enables the full solver suite.

## Features

- **Solves LP, QP and MIP models:** Handles linear, quadratic and mixed-integer linear and quadratic problems, choosing a solver on the server for each type.
- **Named variable results:** Returns the solution state, objective value and each variable's value keyed by variable name.
- **Format translation:** Converts models between LP, MPS and ojAlgo's EBM format through a separate tool.
- **Progress reporting and abort:** Long solves can stream progress to a client that requests it, and ending the call aborts the solve.
- **Self-hosted option:** The same service runs as a public container image on your own infrastructure so models stay in your environment.
- **Licence-based limits:** Free use is restricted to small models and short solves, while a subscription key unlocks larger models and the full solver suite.

## Good for

- Formulating and solving a linear programme from a plain-language description inside an assistant chat
- Checking whether a scheduling or allocation model is feasible and reading off optimal variable values
- Converting an existing MPS model into LP format for easier review
- Running small mixed-integer models from a coding agent during development before moving to a self-hosted deployment

## Install

### Claude Code (Terminal)

```sh
claude mcp add --transport http optimisation-service https://service-api.optimatika.se/optimisation/mcp
```

### Claude Desktop (Settings → Connectors → Add custom connector)

```sh
https://service-api.optimatika.se/optimisation/mcp
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "optimisation-service": {
      "url": "https://service-api.optimatika.se/optimisation/mcp"
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "optimisation-service": {
      "type": "http",
      "url": "https://service-api.optimatika.se/optimisation/mcp"
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "optimisation-service": {
      "serverUrl": "https://service-api.optimatika.se/optimisation/mcp"
    }
  }
}
```

### Grok Bot (Grok Bot desktop app → Connect apps → Add a custom connector)

```sh
https://service-api.optimatika.se/optimisation/mcp
```

Enter a Name (for example "Optimatika Optimisation Service"), choose Type: Remote HTTPS, paste this as the Server URL and choose Add plugin.

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "optimisation-service": {
      "type": "http",
      "url": "https://service-api.optimatika.se/optimisation/mcp"
    }
  }
}
```

## Tools

All 3 tools say they're read-only. From the server's own tool annotations; not verified by us.

- `solve` (says read-only): Solves a linear, quadratic or mixed-integer optimisation model and returns the solution state, the objective value and each variable's value by name. Give the model as text, in LP format unless 'format' says otherwise. For example: Maximize obj: 3 x + 5 y Subject To budget: x + y <= 10 Bounds 0 <= x <= 8 General x End A coefficient goes before its variable with a space, as in 3 x: there is no * fo
- `translate` (says read-only): Rewrites an optimisation model from one file format to another: LP (CPLEX LP format), MPS, or EBM (ojAlgo's own).
- `environment` (says read-only): Reports what a model solved here may be: how large, how long it may run, and which solvers are available.

## FAQ

### Is there a free way to use it?

Yes. Calls to the shared instance without a licence key are free, need no sign-up, and are limited to restricted mode, which allows models of up to 1,000 variables or constraints.

### How do I connect it to Claude Code?

Run claude mcp add --transport http optimisation-service https://service-api.optimatika.se/optimisation/mcp, optionally adding an Authorization Bearer header with a licence key.

### Are submitted models stored?

Models and results are held in memory only on the instance that accepts them, with caches that expire one hour after last access. On the shared instance the server is Optimatika's, so confidential models should not be sent there.

### Which model formats does it accept?

The solve tool accepts LP, MPS or EBM text, and the translate tool converts between those three formats.

### Is Optimatika Optimisation Service free?

Optimatika Optimisation Service has a free plan, with paid plans for more usage or features.

### How do I add Optimatika Optimisation Service to Claude Code?

Run this command in your terminal: claude mcp add --transport http optimisation-service https://service-api.optimatika.se/optimisation/mcp

### What are the best alternatives to Optimatika Optimisation Service?

The closest alternatives to Optimatika Optimisation Service are MCP Z3 Prover, NEOMA P-CAM Physical AI & Ising MCP Server, Algenta MCP Server, pyeconomics and optlens.

## Where this comes from

- Description: written from the project's website and its MCP Registry entry.
- Status and tools: checked automatically by AIPoweredBot on 19 November 2026.
- Install steps: from the official MCP Registry entry.

## Alternatives

- [MCP Z3 Prover](https://ai-powered.com/mcp/mcp-z3-prover): MCP Z3 Prover is an open-source MCP server for developer tools.
- [NEOMA P-CAM Physical AI & Ising MCP Server](https://ai-powered.com/mcp/neoma-p-cam-physical-ai-and-ising-mcp-server): NEOMA P-CAM Physical AI & Ising MCP Server is a freemium MCP server for developer tools.
- [Algenta MCP Server](https://ai-powered.com/mcp/algenta-mcp-server): Algenta MCP Server is a freemium MCP server for developer tools.
- [pyeconomics](https://ai-powered.com/mcp/pyeconomics): pyeconomics is an open-source MCP server for finance and payments.
- [optlens](https://ai-powered.com/mcp/optlens): optlens is an open-source MCP server for developer tools.
- [Local AI MCP](https://ai-powered.com/mcp/local-ai-mcp): Local AI MCP is an open-source MCP server for ai models and llm tooling.
- [onomeo](https://ai-powered.com/mcp/onomeo): onomeo is a freemium MCP server for ai models and llm tooling.
- [Nerdy by Nerd/Noir](https://ai-powered.com/mcp/nerdy-by-nerd-noir): Nerdy by Nerd/Noir is a free MCP server for developer tools.

---
Source: https://ai-powered.com/mcp/optimatika-optimisation-service
