Agent setup

Run the agent in OpenClaw

OpenClaw MCP support is rolling out. Add the SpendSignoff server to the OpenClaw MCP config, authorize an ad account read-only, and approve drafts in the SpendSignoff app.

Your AI can read and draft — it can never spend without your approval.

OpenClaw support is rolling out

OpenClaw is an open agent client. SpendSignoff support is on the way: the plan is a remote MCP entry registered over OAuth 2.1 with the mcp.read and mcp.draft scopes only.

The intended config is below. Whatever the client surface looks like, the scope boundary holds — the client reads and drafts, approval lives in the SpendSignoff app.

Intended config

Add the SpendSignoff server to the OpenClaw MCP config:

mcp.json

{
  "mcpServers": {
    "spendsignoff": {
      "url": "https://mcp.spendsignoff.com/mcp"
    }
  }
}

Setup steps

Add the MCP server entry

Add the spendsignoff entry to your OpenClaw MCP config and restart the client. The OAuth 2.1 handshake runs on first connection, requesting mcp.read and mcp.draft.

Authorize an ad account

Open the SpendSignoff dashboard and connect a Google Ads or Meta account. The agent can read the account immediately; write operations produce drafts that wait for your approval.

Review and approve your first draft

Ask the agent to propose a change. It returns a before-and-after diff via propose_change. Confirm it in the SpendSignoff app approval queue using the two-step control.

Read and draft only — the agent cannot spend your budget

The SpendSignoff Performance Marketing Agent holds mcp.read and mcp.draft scopes only. It reads account data and produces change drafts. It cannot push a change live. Every live change requires a two-step Approve and Confirm in the SpendSignoff app and is recorded in a KMS-signed append-only audit log.

Run in Claude Code

Claude Code is live today — start there while OpenClaw rolls out.

SpendSignoff documentation

OpenClaw support is rolling out

Intended config

Setup steps