MCP & CLI

Ask your AI agent about your keywords.

Veldo includes an MCP server, so Claude Code, Claude Desktop, or any Model Context Protocol client can read your keyword ranks, Apple Ads spend, and verdicts and reason about them. "Which keywords am I losing money on?" becomes a question you ask in your editor, answered from the database on your Mac.

Setup

For Claude Code, one command:

claude mcp add veldo -- /Applications/Veldo.app/Contents/Helpers/veldo mcp

For Claude Desktop or another MCP client, add this to its MCP config:

{
  "mcpServers": {
    "veldo": {
      "command": "/Applications/Veldo.app/Contents/Helpers/veldo",
      "args": ["mcp"]
    }
  }
}

What the agent can see

Tool	Returns
`list_apps`	Your tracked apps
`list_keywords`	Keyword rows with current rank and change for an app
`get_advisor`	Rank, 30-day spend and installs, and the verdict for every keyword

Read-only, and local

The server answers from the Veldo database on your Mac over stdio. It exposes no write operations: an agent cannot pause campaigns, change bids, or edit keywords. And because the data never leaves your machine, connecting an agent adds no new party to your data.

Prefer plain scripts?

The same binary is a CLI with JSON output, ready for jq, cron, or CI:

veldo apps                  # tracked apps
veldo keywords <appID>      # ranks + deltas
veldo advisor <appID>       # rank + spend + verdict per keyword

Join the waitlist