# tokenometer MCP server

> Calculate LLM costs, count tokens, and benchmark latency across multiple providers like Claude, GPT-4o, and Gemini.

- Repository: https://github.com/faraa2m/tokenometer
- Page: https://hostedmcp.org/servers/faraa2m/tokenometer
- Transport: —
- Stars: 1
- License: MIT

## Environment variables

| Name | Required | Secret | Description |
|---|---|---|---|
| `ANTHROPIC_API_KEY` | no | no |  |

## FAQ

### Is the token count always exact?

Not always; it provides an 'Approx' flag to indicate if the count is a proxy or an exact match from the provider.