Take this research further with AI

1. Connect the MCP server

This lets an agent list, search and read every entry in the guide directly.

https://www.californiapeople.org/api/mcp

• Claude (web, desktop or mobile): Settings → Connectors → Add custom connector, then paste the URL above. No sign-in needed.
• Claude Code:

claude mcp add --transport http ca-ballot-guide https://www.californiapeople.org/api/mcp

• Other MCP clients (Cursor, VS Code and others):

{
  "mcpServers": {
    "ca-ballot-guide": {
      "url": "https://www.californiapeople.org/api/mcp"
    }
  }
}

Tools: list_entries, get_entry, search, and get_all.

2. Copy a prompt

Every proposition, office and local page has a Copy for AI button. It copies that page's research and sources, plus links the agent can follow to explore the rest of the guide. To start from the whole guide instead, paste this into any chatbot and add your question at the end:

You are helping a California voter research the November 3, 2026 ballot. Start from the California Ballot Guide, a sourced research guide to every proposition, statewide office and San Bernardino County local race.

Index of every entry: https://www.californiapeople.org/llms.txt
Everything in one file: https://www.californiapeople.org/llms-full.txt

Read the entries relevant to my question, check key claims against their linked sources, and dig deeper where the evidence is thin: track records, who gets the money, cost per outcome, conflicts of interest. Stay factual and show both sides.

My question: 

3. Raw files

• llms.txt: an index of every entry, with links
• llms-full.txt: the entire guide in one file
• data/all.json: everything as JSON
• /md/<id>.md: one entry as Markdown, for example md/prop-40.md

Your picks on this site stay in your browser and are never part of these files.