Caliper
by Community
Run your real agent with and without your skills, MCPs, and rules. See which ones actually help, and what they cost in tokens. Supports Claude Code, Codex, Pi, and Hermes.
OSS
Caliper
Added 16 Oct 2026
Overview
Caliper runs your real agent with and without your skills, MCPs, and rules, so you can see which ones actually help and what they cost in tokens. It supports Claude Code, Codex, Pi, and Hermes. The tool is implemented in Python.
Best for
Best for
Python-based agent builders who want empirical answers about what skills actually pay off
Use cases
- Compare agent runs with and without specific skills
- Measure token costs of MCPs and rules
- Evaluate agent configurations across Claude Code, Codex, Pi, and Hermes
Notes
212 stars on GitHub. Last updated 2026-10-16. Licensed MIT.
Pros
- Uses real agent runs rather than synthetic tests
- Shows token cost impact of each addition
- Supports multiple agent frameworks
Cons
- Community project with limited adoption (212 stars)
- Requires Python environment to run
- No built-in visualization or dashboard mentioned
Get the free Developer’s Field Guide
A 27-page field guide to the AI coding workflow with Claude. Claude Code, MCP servers, the prompt patterns that work, and what to delegate. Free.
Enter your work email. We send it straight over, plus a few short notes worth knowing. Unsubscribe any time.