# Roadie API & MCP Servers

> An introduction to the Roadie API and MCP servers for programmatic access to your Roadie instance.

*Published: 2022-11-15*


## Overview

Roadie provides both a REST API and Model Context Protocol (MCP) servers to read and write information from a Roadie instance. These enable you to integrate Roadie with your existing tools, automate workflows, and build custom integrations.

## Roadie API

The Roadie API provides programmatic access to your Backstage catalog and other Roadie features. You can use it to automate catalog management, integrate with CI/CD pipelines, or build custom tooling.

Some uses of the API include:

- **Catalog management** - Create, update, and delete entities in your catalog from external systems
- **Entity sets** - Manage batches of entities idempotently, perfect for syncing resources from internal systems
- **Tech Insights** - Access fact data and scorecard results programmatically

To get started with the API, you'll need to [generate an API token](/docs/api/authorization/).

### API Base URL

```
https://api.roadie.so/api/
```

### Example: List catalog entities

```bash
curl \
  -X GET \
  -H 'Accept: application/json' \
  -H "Authorization: bearer ${ROADIE_API_TOKEN}" \
  https://api.roadie.so/api/catalog/entities
```

For more details, see the [API Authorization](/docs/api/authorization/) documentation and the specific API documentation for [Catalog](/docs/api/catalog), [Tech Insights](/docs/api/techinsights), [Templates](/docs/api/templates), and [Plugins](/docs/api/plugins).

## MCP Server(s)

You can also connect to your Roadie instance via an LLM client of your choice (provided it supports third party MCP servers). [Model Context Protocol (MCP)](https://modelcontextprotocol.io/introduction) enables AI assistants to interact with your Backstage catalog using structured data.

Some uses of the MCP servers include:

- **AI-powered catalog exploration** - Ask natural language questions about your software catalog and/or allow Agents to access rich metadata about your software.
- **Automated scaffolding** - find, validate, and execute scaffolder templates
- **Security and compliance insights** - Query vulnerability data, branch protection status, and compliance metrics
- **Documentation search** - Search and retrieve TechDocs content across your catalog
- **Catalog decoration** - Manage entity fragments and decorators via AI assistants

### Available MCP Servers

| Server                                                           | Description                                                           | Endpoint                                               |
| ---------------------------------------------------------------- | --------------------------------------------------------------------- | ------------------------------------------------------ |
| [API Docs Query](/docs/api/roadie-mcp/api-docs-query/)           | Discover and retrieve API documentation and specifications            | `https://api.roadie.so/api/mcp/v1/api-docs-query`      |
| [Backend Config](/docs/api/roadie-mcp/backend-config/)           | Manage proxy settings and secrets                                     | `https://api.roadie.so/api/mcp/v1/backend-config`      |
| [Catalog Decorators](/docs/api/roadie-mcp/catalog-decorators/)   | Manage catalog entity decorators and fragments                        | `https://api.roadie.so/api/mcp/v1/catalog-decorators`  |
| [Rich Catalog Entity](/docs/api/roadie-mcp/rich-catalog-entity/) | Access catalog entity data, relationships, and documentation          | `https://api.roadie.so/api/mcp/v1/rich-catalog-entity` |
| [Scaffolder](/docs/api/roadie-mcp/scaffolder/)                   | Find, validate, and execute scaffolder templates                      | `https://api.roadie.so/api/mcp/v1/scaffolder-use`      |
| [Tech Insights Facts](/docs/api/roadie-mcp/tech-insights-facts/) | Access operational metrics, security data, and compliance information | `https://api.roadie.so/api/mcp/v1/tech-insights-facts` |

### Supported AI Tools

MCP servers work with popular AI development tools including:

- **VS Code with Copilot**
- **Cursor IDE**
- **Claude Desktop**

For detailed setup instructions, see the [Roadie MCP Getting Started](/docs/api/roadie-mcp/) documentation.

## Authentication

Both the REST API and MCP servers use the same authentication mechanism. You'll need to generate an API token from your Roadie instance:

1. Go to **Administration → Account**
2. Navigate to the **Roadie API Access** section
3. Add a token description and click **Generate Token**
4. Store the token securely

See [API Authorization](/docs/api/authorization/) for more details.