Labsco
deepgram logo

setup-mcp

★ 13

by deepgram · part of deepgram/skills

Set up the Deepgram MCP server for your AI coding tool. Checks whether the Deepgram CLI (dg/deepctl) is installed: if so, uses the local CLI MCP server (dg mcp) for full tool access; otherwise offers the hosted documentation MCP or suggests installing the CLI. Use whenever someone wants to install Deepgram's agentic tools, set up the MCP server, or connect their editor to Deepgram.

🔥🔥✓ VerifiedFreeQuick setup
#voice#git
🧩 One of 7 skills in the deepgram/skills package — works on its own, and pairs well with its siblings.

This is the playbook your agent receives when the skill activates — you don't need to read it to use the skill, but it's here to audit before installing.

Install the Deepgram MCP Server

You are setting up Deepgram MCP integration for the user. Follow these steps:

Step 1: Check for the Deepgram CLI

Run dg --version (or deepctl --version, or where dg on Windows) to check if the Deepgram CLI is installed.

  • If found: use the local CLI MCP (dg mcp) — this gives full access to Deepgram tools including transcription, text-to-speech, project management, and more.
  • If not found: use the hosted documentation MCP and offer to install the CLI.

Step 2: Detect the environment

Determine which AI coding tool the user is running. Check for:

  • Claude Code — look for a .claude/ directory in the project or user home
  • Cursor — look for a .cursor/ directory in the project root
  • Windsurf — look for a .windsurf/ directory in the project root

If multiple are detected, or none are detected, ask the user which tool they want to configure.

Step 3: Ask about scope

Ask the user whether they want the MCP server configured:

  • For this project only (recommended for team repos)
  • Globally (available in all projects)

Step 5: Confirm

After installation:

  • Claude Code — run /reload-plugins to activate immediately, no restart needed.
  • Cursor / Windsurf / Other — the user may need to restart or reload their tool.

Then tell the user:

The Deepgram MCP server is now configured.

If using the CLI MCP (dg mcp), add:

Available tools include transcription, text-to-speech, project management, and usage queries. Try asking to transcribe an audio file or convert text to speech.

If using the hosted docs MCP, add:

Your tool can now query Deepgram's full documentation directly — try asking about API parameters, voice agents, or model capabilities.

Link them to Deepgram Agentic Tools for more details.