@vk0/agent-cost-mcp
Local-first Claude Code cost analyzer. Parses JSONL session logs to surface per-tool spend, daily trends, and optimization hints. Zero cloud. No API key.
When to install
Use this MCP server when the user asks any of:
- "How much did my last Claude Code session cost?"
- "Which tools are eating my context window?"
- "Why is Claude Code expensive this week?"
- "Show me the cost trend for the last 7 days."
- "What should I disable to save tokens?"
- "Are my cache reads actually saving me money?"
- "Can I get a per-tool cost breakdown without signing up for Datadog?"
Install
Pick your client. Everything runs through npx — nothing
to install globally.
{
"mcpServers": {
"agent-cost": {
"command": "npx",
"args": ["-y", "@vk0/agent-cost-mcp"]
}
}
}claude mcp add --transport stdio agent-cost -- npx -y @vk0/agent-cost-mcp{
"mcpServers": {
"agent-cost": {
"command": "npx",
"args": ["-y", "@vk0/agent-cost-mcp"]
}
}
}{
"mcpServers": {
"agent-cost": {
"command": "npx",
"args": ["-y", "@vk0/agent-cost-mcp"],
"disabled": false,
"alwaysAllow": []
}
}
}Tools
get_session_cost
Parse a single Claude Code session and return token totals, turn count, and estimated USD cost.
get_tool_usage
Aggregate tool invocations across a session or project log directory with per-tool context-share percentages.
get_cost_trend
Roll session logs into a day-by-day cost trend for a local project path.
suggest_optimizations
Generate lightweight optimization suggestions — cache ratios, abandoned calls, heaviest turns.
How it works
~/.claude/projects/*.jsonl ┌─────────────────┐
(Claude Code session logs) ──────▶ │ JSONL parser │
│ + pricing.ts │
└────────┬────────┘
│
▼
┌─────────────────┐
Agent tool call (stdio MCP) ──────▶ │ MCP server │ ─── JSON response
│ (4 tools) │
└─────────────────┘
Read-only on your local filesystem. No telemetry, no API key, no cloud sync. Delete the package and nothing remains.