Launching today

MCPSpend

Launching today

Know what your AI agents really cost

3 followers

Know what your AI agents really cost

3 followers

Visit website

MCPSpend is the observability proxy for Model Context Protocol tools. Track tokens, attribute spend per team and customer, and ship AI to production with audit-grade visibility.

Free

Launch tags:Developer Tools•Artificial Intelligence•Data & Analytics

Launch Team / Built With

Layout Agent 2.0 — Thinks before it builds

Thinks before it builds

Promoted

Maker

📌

i PH 👋 I'm Andrei, solo founder out of Bucharest. I built MCPSpend after watching my Cursor + Claude Desktop bills compound silently for 3 months — no breakdown by tool, no breakdown by project, just one big surprise number at the end of the cycle. If you use any MCP server (Playwright, filesystem, GitHub, Figma-remote, you name it), MCPSpend is the layer between your AI client and those servers. Every tool call is logged with cost, latency, tokens, success/fail. Dashboard, CSV export, budget alerts — everything you'd expect from observability tools, just built for the agent era. Install is one command: npx --yes @mcpspend/proxy@latest init --key mcps_live_xxx It auto-detects Claude Desktop, Cursor, Windsurf, VS Code, and Claude Code on your machine and wraps every MCP server you have configured. No code changes, no SDK swap, fire-and-forget — if our API is down, your agent keeps working. Free tier: 25k tool calls/month, no credit card. Paid plans start at $29/mo, yearly with 2 months free. What I want from you today: 1. Try it (npx, 60 seconds) 2. Tell me where it broke / what was confusing 3. What's the most expensive MCP tool you've ever seen? — I'm building a small leaderboard with the funniest answers.

Report

15h ago

Maker

Seeto.ai knows better 💪

Report

3h ago

Forum Threads

p/mcpspend

•

15h ago

After tracking 30 days of MCP tool calls, here's what I learned

I spent the last 30 days logging every MCP tool call I made across
Cursor, Claude Code, and Claude Desktop. Some surprises: 1. Read tool calls dominate by volume but tokens are tiny they're cheap. The cost spike comes from a handful of big calls (web search, image processing, large file reads). 2. ~14% of tool calls fail or timeout. They still cost something on the LLM side because the model already burned input tokens. 3. There's no native way in Cursor/Claude/Windsurf to see which project or which session caused the spend. You only see totals. 4. Per-agent attribution flips assumptions: "the Claude Code in my dev folder" vs. "the Cursor in my marketing folder" had a 7x cost delta for the same number of files touched. I turned the logger into a product (MCPSpend) and we're launching on
PH tomorrow. Free tier 25k calls/month if anyone wants to test:
npx --yes @mcpspend/proxy@latest init --key mcps_live_xxx Curious anyone else tracking their MCP spend manually? What's
your stack?

View all