CLI commands

#

DocStar CLI — Available Commands

#

Overview

DocStar CLI allows users to interact with APIs documented and published through DocStar. The CLI can discover API modules from a DocStar documentation collection, install them locally, execute API endpoints as commands, and expose those APIs through MCP.

#

1. Initialize a Documentation Collection

#

Basic

docstar-cli init <docs-domain>

Discovers the published modules from a DocStar documentation site and installs them locally.

#

With a Collection ID

docstar-cli init <domain> --collectionId <id>

Use this when the collection needs to be explicitly identified.

#

Example

docstar-cli init docs.msg91.com
#

2. List Installed Collections

docstar-cli list

Displays the DocStar collections currently installed on the system.

#

3. List Modules

docstar-cli list <cli-name>

Lists the modules available inside an installed collection.

#

Example

docstar-cli list msg91
#

4. List Endpoints in a Module

docstar-cli list <cli-name> <module>

Lists the available API endpoints/commands inside a specific module.

#

Example

docstar-cli list msg91 slack

This can show commands such as:

send-message-1
#

5. Execute an API Endpoint

The general command format is:

docstar-cli <cli-name> <module> <endpoint>
#

Example

docstar-cli msg91 slack send-message-1 --message "hello from the CLI"

The CLI translates the command into the API request defined by the endpoint’s DocStar configuration.

The endpoint definition can contain information such as:

  • HTTP method

  • API URL

  • Base URL

  • Headers

  • Request body

  • Required parameters

  • Optional parameters

#

6. Interactive Parameters

An endpoint can also be executed without providing all parameters directly:

docstar-cli msg91 slack send-message-1

The CLI can interactively ask the user for the required parameters.

This makes API usage possible without manually constructing HTTP requests.

#

7. Start the MCP Server

docstar-cli mcp

Starts the DocStar CLI as an MCP server.

Installed API endpoints can then be exposed as MCP tools for AI clients.

For example, an endpoint such as:

msg91
└── slack
    └── send-message-1

can be exposed as an MCP tool similar to:

msg91__slack__send-message-1

An AI agent can then call the API through the MCP tool instead of manually constructing an HTTP request.

#

Command Structure

The CLI has two types of commands.

#

Built-in Commands

These are implemented directly by the CLI:

init
list
mcp
#

Dynamically Generated API Commands

These come from the API documentation published through DocStar:

<collection>
└── <module>
    └── <endpoint>

For example:

docstar-cli msg91 slack send-message-1

Here:

Part

Meaning

msg91

DocStar collection / CLI name

slack

API module

send-message-1

API endpoint

--message

Endpoint parameter

#

Typical Workflow

A user would typically follow this flow:

1. Initialize a DocStar collection
          ↓
2. CLI discovers published API modules
          ↓
3. Modules/endpoints are installed locally
          ↓
4. User lists available APIs
          ↓
5. User executes an API endpoint
          ↓
6. CLI builds and sends the API request

Example:

docstar-cli init docs.msg91.com

docstar-cli list msg91

docstar-cli list msg91 slack

docstar-cli msg91 slack send-message-1 --message "Hello"

For AI/agent usage:

docstar-cli mcp

This exposes the installed API capabilities through MCP.

#

Key Design Idea

The CLI does not need to hardcode every customer’s API.

Instead:

DocStar API Documentation
        ↓
Published API specification
        ↓
docstar-cli init
        ↓
Local API definitions
        ↓
Generated CLI commands

Therefore, if another company publishes its API through DocStar, the same CLI can potentially generate commands for that company’s APIs without creating a separate CLI implementation.

Built with DocStar