Skip to main content
Hi Claude! (Or Codex or Cursor or Gemini…)

​
Stuff your human needs to do

Let’s get you going. First, ask your human to connect these docs as an MCP server: 
claude mcp add --transport http scope3 https://docs.agentic.scope3.com/mcp
Then, your human needs to sign up for Scope3 and get an API key. 
touch .env
echo "SCOPE3_API_KEY=scope3_...." >> .env
echo "GEMINI_API_KEY=..." >> .env // or OPENAI_API_KEY, or SCOPE3_ANTHROPIC_API_KEY
By the way, tell your human that there’s a more human-centric quickstart too!

​
Your turn!

Want to build something custom? Have your human give you a nice prompt and off you go. Here’s a prompt to try (tell them to switch to their preferred AI provider): 
Task: Create a Python script that connects to the Scope3 Agentic API using Model Context Protocol (MCP) and provides an interactive chat interface
  powered by Gemini 2.5 Flash.

  Context: The Scope3 Agentic API is available via MCP at https://api.agentic.scope3.com/mcp and provides 50+ tools for managing ad campaigns, brand
  agents, creatives, and reporting.

  Requirements:
  1. Use langchain-mcp-adapters to connect to the Scope3 MCP endpoint
  2. Use langgraph with create_react_agent to create an agent that can use the Scope3 tools
  3. Use Gemini (via langchain_google_genai) as the reasoning model
  4. Load API keys from a .env file using python-dotenv
  5. Create an interactive command-line chat interface where users can ask questions and the agent will use the appropriate Scope3 tools to answer