
Local cost attribution and guardrails for Claude Code AI sessions.
Visit Agent Cost McpAgent Cost MCP is a local-first Model Context Protocol (MCP) server that provides AI session cost forensics, usage attribution, budget guardrails, and cost anomaly detection for Claude Code operators and developers. It parses local Claude Code JSONL session logs to attribute spending by tool, model, and subagent tree, forecast future costs, detect anomalies, set budget caps, and push signed spend alerts—all without uploading data to the cloud or requiring API keys. Ideal for developers and operators who want transparent, machine-readable insights into which branches and tools are driving token consumption, plus pre-spend guardrails and actionable optimization suggestions.
Visit Agent Cost Mcp's official website for product details and getting started.
Comprehensive guides and API references for implementing Agent Cost MCP.
Comprehensive guidelines and API documentation for setting up and using Agent Cost MCP.
A community-driven space for reporting bugs, requesting features, and discussing Agent Cost MCP.