Agentic developer tools
Deepgram provides multiple ways to embed Deepgram knowledge directly into AI-powered developer tools: a CLI with a built-in MCP server, a docs MCP server for querying documentation, and a skills marketplace with structured reference material your tools can load on demand.
Deepgram CLI
Section titled “Deepgram CLI”The dg CLI gives you full access to Deepgram APIs from your terminal — speech-to-text, text-to-speech, voice agents, and more — with 28 commands and a built-in MCP server.
Install
Section titled “Install”macOS / Linux:
curl -fsSL https://deepgram.com/install.sh | shWindows (PowerShell):
iwr https://deepgram.com/install.ps1 -useb | iexPackage managers:
pip install deepctl # pip
uv tool install deepctl # uv
pipx install deepctl # pipxBuilt-in MCP server
Section titled “Built-in MCP server”The CLI includes a built-in MCP server that gives your AI coding tools direct access to Deepgram APIs:
dg mcp # Start MCP server (stdio)
dg mcp --transport sse --port 8000 # SSE transportAdd to your editor’s MCP configuration:
{
"mcpServers": {
"deepgram": {
"type": "stdio",
"command": "dg",
"args": ["mcp"]
}
}
}For full CLI documentation, visit the CLI documentation on developers.deepgram.com, or explore cli.deepgram.com for the quick reference.
Docs MCP server
Section titled “Docs MCP server”The Deepgram docs MCP server lets AI coding tools query the full Deepgram documentation. It is available over HTTP at two endpoints:
| Transport | HTTP |
| URL | https://api.dx.deepgram.com/kapa/mcp |
| Alternative URL | https://deepgram.mcp.kapa.ai |
Claude Code
Section titled “Claude Code”Add the server with a single command:
claude mcp add deepgram-docs --scope project --transport http https://api.dx.deepgram.com/kapa/mcpCursor, Windsurf, and other MCP clients
Section titled “Cursor, Windsurf, and other MCP clients”Add the following to your MCP configuration file (for example, .cursor/mcp.json or .windsurf/mcp.json):
{
"mcpServers": {
"deepgram-docs": {
"type": "http",
"url": "https://api.dx.deepgram.com/kapa/mcp"
}
}
}Once connected, your AI tool can answer questions about any Deepgram API, SDK, or feature using the official documentation as its source.
Skills marketplace
Section titled “Skills marketplace”The Deepgram skills repository contains agent-agnostic skill files that AI coding tools load to understand specific parts of the Deepgram platform. Each skill is a folder of plain Markdown that any tool can consume.
Available skills
Section titled “Available skills”| Skill | Description |
|---|---|
api |
Full API reference for all Deepgram REST and WebSocket APIs, generated from OpenAPI and AsyncAPI specs |
docs |
Documentation navigator with topic-by-topic pointers and MCP server setup instructions |
starters |
Ready-to-run demo apps across 13 frameworks and 7 features |
setup-mcp |
Set up the Deepgram MCP server for your AI coding tool |
Quick setup (Claude Code)
Section titled “Quick setup (Claude Code)”Register the repository as a plugin marketplace:
/plugin marketplace add deepgram/skillsInstall the Deepgram plugin:
/plugin install deepgram@deepgram-agent-skillsThis gives you the following slash commands: /deepgram:api, /deepgram:docs, /deepgram:starters, and /deepgram:setup-mcp.
Deepgram provides official SDKs so you can integrate speech-to-text, text-to-speech, voice agents, and audio intelligence into your application in the language of your choice.
Python

Install with pip install deepgram-sdk
JavaScript / TypeScript

Install with npm install @deepgram/sdk
Go

Install with go get github.com/deepgram/deepgram-go-sdk
.NET

Install with dotnet add package Deepgram
Java

Install with Maven or Gradle — com.deepgram:deepgram-java-sdk
Rust
Install with cargo add deepgram
For a full breakdown of which SDK supports which API feature, see the SDK feature matrix.