$ loading_
Track multi-provider LLM API spend locally and expose it via MCP.
Copy the install command and let the AI configure it · recommended for beginners
Please install the "LLM Usage" MCP server from askskill: Run: claude mcp add 'io-github-zhaoyue722-llm-usage-mcp' -- uvx llm-usage-mcp
Read the last 7 days of API spend from LLM Usage, summarize by provider and model, and identify the highest-cost items.
A cost summary grouped by provider and model, plus highlights of the most expensive items.
Check whether recent LLM API spending shows abnormal spikes, and list the time periods and related models that may have surged.
A list of anomalous periods, involved models or providers, and a brief explanation of the cost changes.
Using spend data from LLM Usage, tell me which models or providers currently cost more so I can adjust model choices for upcoming coding tasks.
A spend-based cost comparison that can inform later coding-agent model selection.
When developers use multiple LLM providers, they can use this tool to capture API spend in one place and let coding agents read that cost data directly. This makes it easier to understand usage costs across models.
When using MCP-enabled coding agents, teams can use this tool as a cost data source so the agent can reference existing LLM spend records during analysis or recommendations. It fits workflows that balance budget and productivity.
For teams that prefer to manage model usage data locally, this tool can capture LLM API spend locally and expose it through MCP to other development workflows. It suits spend tracking scenarios that value local control.
It is a local-first, multi-provider MCP tool that captures LLM API spend and makes that data available to coding agents.
Yes. The original description explicitly says it is a multi-provider tool. However, the provided materials do not specify which providers are supported; see the source repository.
We only know that it exposes data to coding agents through MCP. The provided materials do not specify installation steps, runtime requirements, or key requirements; see the source repository.
Offload bounded text tasks from coding agents to local or cheaper LLMs.
Route coding tasks across local and remote LLMs with benchmarking and code search.
Run Llama models locally for private, offline AI assistance.
Delegate low-risk tasks to a cheaper model with main-agent review.
Attribute multi-tenant LLM usage costs for chargeback billing.
Run local multi-model deliberation and synthesis on Mac for AI coding workflows.