# rocketride-org/rocketride-server [Health: Active]

**Category:** 🛠️ Other Tools and Integrations  
**Repository:** https://github.com/rocketride-org/rocketride-server  
**GitHub Stars:** 17753  
**Views:** 6  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/rocketride-org-rocketride-server

## Description
MCP server that exposes RocketRide AI pipelines as tools for Claude, Cursor, and Windsurf. Self-hosted, open-source pipeline tool with multi-LLM support.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `uvx` (confidence: high):

```json
"mcpServers": {
  "rocketride-server": {
    "command": "uvx",
    "args": ["rocketride"]
  }
}
```

## Documentation

## What rocketride-org/rocketride-server does

rocketride-org/rocketride-server makes RocketRide AI pipelines available to MCP clients as callable tools. Instead of embedding each workflow directly in an assistant integration, a client can invoke a pipeline exposed through RocketRide. This is suited to teams that already build workflows with RocketRide or want agents to access repeatable AI and data-processing operations.

RocketRide is an open-source pipeline builder and runtime designed for AI and machine-learning workloads. Its pipeline system covers LLM providers, vector databases, OCR, named-entity recognition, embeddings, chunking, and related processing tasks. The MCP server provides the connection point between those workflows and supported AI assistants.

## How it works

Pipelines are composed in RocketRide and stored as portable JSON definitions. They can be created visually in VS Code, tested, observed, and executed by RocketRide’s runtime. Through MCP, the resulting workflows are presented to an assistant as tools that can be called during an interaction.

Execution remains tied to the RocketRide environment hosting the pipelines. The platform supports local operation and self-hosted deployment, so teams can run workflows on their own infrastructure rather than sending pipeline execution through a hosted service. RocketRide also supports Python and TypeScript SDKs for applications that need to invoke pipelines outside an MCP client.

The underlying runtime supports multiple LLM providers and other pipeline nodes. This allows an exposed workflow to combine model calls with retrieval, document processing, or other configured stages, depending on how the pipeline was built.

## Setup and configuration

The server is intended for self-hosted use with RocketRide. RocketRide can run locally, and its MCP documentation provides the connection path for exposing pipelines to clients. The project links to setup documentation at [docs.rocketride.org/connect/mcp/http](https://docs.rocketride.org/connect/mcp/http).

Client support called out for this server includes Claude, Cursor, and Windsurf. The broader RocketRide project also provides a VS Code extension for designing and running pipelines, but pipeline authoring and MCP client connectivity are separate parts of the workflow.

## Tools and capabilities

- Expose RocketRide pipelines as MCP-callable tools.
- Run reusable AI workflows from Claude, Cursor, or Windsurf.
- Use pipelines that combine multiple LLM providers and data-processing nodes.
- Work with workflows involving vector databases, OCR, NER, embeddings, and chunking.
- Keep pipeline definitions portable as JSON files.
- Execute workflows on local or self-managed infrastructure.

RocketRide advertises more than 100 pipeline nodes, spanning more than 15 LLM providers and nine vector databases. It also includes support for multi-agent workflows through CrewAI and LangChain at the platform level.

## Limitations and notes

The server exposes capabilities defined by the RocketRide pipelines available to it; it is not a general-purpose collection of unrelated MCP tools. The useful operations, model providers, databases, and processing stages therefore depend on the pipelines and nodes configured in the RocketRide environment.

RocketRide is designed for self-hosted operation, which places runtime and deployment responsibilities on the operator. The project is MIT-licensed and supports local, Docker, on-premises, and RocketRide Cloud deployment options for the broader platform.

_Full upstream README: https://allmcps.com/mcp/rocketride-org-rocketride-server/readme_