<p>The CFO agent analyzes AI token spend and recommends optimizations.
It connects to OpenRouter's MCP server for real usage and pricing data.</p>
<h2 id="goal">Goal</h2>
<p>Optimize AI spend by tracking token usage across models, identifying
expensive patterns, and recommending model or workflow changes that
reduce cost without sacrificing quality.</p>
<h2 id="tools">Tools</h2>
<ul>
<li><strong>OpenRouter MCP</strong> — usage data (tokens, cost per model) and model
pricing information</li>
<li><strong>Read</strong> — access project files and wiki for context</li>
</ul>
<h2 id="subagents">Subagents</h2>
<table>
<thead>
<tr>
<th>Subagent</th>
<th>Status</th>
<th>Role</th>
</tr>
</thead>
<tbody>
<tr>
<td>Cost Tracker</td>
<td>Active</td>
<td>Produce spend reports, flag anomalies</td>
</tr>
</tbody>
</table>
<h2 id="metrics">Metrics</h2>
<ul>
<li>Daily/weekly/monthly token spend by model</li>
<li>Cost per task type (blog writing, code review, research)</li>
<li>Model pricing changes that affect the budget</li>
<li>Spend anomalies (unexpected spikes)</li>
</ul>
<h2 id="invocation">Invocation</h2>
<pre><code class="language-bash"># Claude Code
claude --agent cfo

# Example prompts
# "What did I spend on AI this week?"
# "Which model is costing me the most per token?"
# "Are there cheaper models I could use for research tasks?"
</code></pre>


Chief Financial Officer agent. Analyzes AI token spend via OpenRouter, identifies cost optimization opportunities.

Kyle Pericak

"It works in my environment"

CFO Agent

Goal

Tools

Subagents

Metrics

Invocation