Agent Cost MCP Server runs on your own machine — the client starts it, so there is no endpoint to ping. 921 installs a week from pypi. Last commit 5 Nov 2026.
Prices LLM/agent runs per successful task, with spend recording and budget status.
Today is the operative word: we check Agent Cost MCP Server every 15 minutes and re-read its code on every release. Watch it and you find out the day that stops being true.
This server runs on your own machine — install it with the package manager and the client starts it for you. Package name taken from the official registry entry.
claude mcp add agent-cost-mcp-server -- uvx agent-cost-mcp-server
{
"mcpServers": {
"agent-cost-mcp-server": {
"command": "uvx",
"args": [
"agent-cost-mcp-server"
]
}
}
}
[mcp_servers.agent-cost-mcp-server]
command = "uvx"
args = ["agent-cost-mcp-server"]
{
"mcpServers": {
"agent-cost-mcp-server": {
"command": "uvx",
"args": [
"agent-cost-mcp-server"
]
}
}
}
{
"servers": {
"agent-cost-mcp-server": {
"type": "stdio",
"command": "uvx",
"args": [
"agent-cost-mcp-server"
]
}
}
}
Optional settings from the registry entry, left out of the block above:
AGENT_COST_LEDGER_PATH (Optional path for the local spend-ledger file used by record_spend / budget_status).
LLM provider intelligence: recommendations, pricing, and status
The budget app your AI agent can use. Check budgets and log spending, with no bank logins.
Rent GPUs at live prices and serve any Hugging Face model, with spend caps for AI agents.
Feedback, roadmap, agent task execution with review, changelog, status pages and error logs.
Agent spending mgmt, budget tracking, ROI. Zero Core Budget.
Connect AI agents to Process Street workflows, tasks, runs, data sets, and operational records.
Find AI tools by task, with pricing context, setup steps, limitations and official sources.
Read public web pages and run goal-based browsing tasks with Rove's sandboxed browser agent.
Answers built from our own checks of this server.