Build with AI

This product is not supported for your selected Datadog site. ().

Datadog supports coding agent workflows through the Datadog MCP server, the Pup CLI, and Claude Code skills. Use them to investigate your Agent Observability data and iterate on your LLM application without leaving your development environment.

Get started

Install the skills

npx skills add datadog-labs/agent-skills/agent-observability --full-depth -y

Choose a data backend

The skills read your Agent Observability data through either the Datadog MCP server or the Pup CLI. Set up one of them. Each skill detects the MCP server at startup and falls back to the Pup CLI when the MCP server is unavailable.

Option A: Datadog MCP server

Connect the MCP server to your Claude Code session:

claude mcp add --scope user --transport http datadog-llmo-mcp \
  '?toolsets=llmobs,core'

To run the MCP server as a local binary instead, install the Datadog MCP CLI:

curl -sSL https://coterm.datadoghq.com/mcp-cli/install.sh | bash
datadog_mcp_cli login
claude mcp add datadog --scope user -- ~/.local/bin/datadog_mcp_cli

This product is not supported for your selected site ().

For the full setup, including authentication options and the available tools, see MCP Server.

Install the Pup CLI with Homebrew (macOS/Linux) and authenticate:

brew tap datadog-labs/pack
brew install datadog-labs/pack/pup
pup auth login

For other installation methods, supported commands, and authentication options, see Pup CLI.

Build with the Datadog MCP server and skills: