# token-optimizer MCP server

> Reduces token consumption by converting large command logs into compact, actionable summaries using local or OpenRouter LLMs.

- Repository: https://github.com/lefos13/token-optimizer
- Page: https://hostedmcp.org/servers/lefos13/token-optimizer
- Transport: —
- Stars: 1
- License: Apache-2.0

## FAQ

### How does Token Optimizer handle sensitive data?

Remote requests are redacted before leaving your machine to strip secret values, and raw logs stay stored locally in your workspace.

### What command profiles are available for security?

You can choose between 'standard' for common build/test scripts, 'safe' for explicitly allowed commands, or 'unrestricted' for trusted repositories.

### Can I use my own local LLM?

Yes, you can point the tools at an OpenAI-compatible local endpoint to ensure no data is sent remotely.

### What happens if the LLM provider fails?

The server will not invent a result; it maintains authoritative exit codes and reports an honest 'uncertain' verdict.