Observe, Optimize & Govern Your AI Coding Fleet

Actively publishing live token & cost telemetry for Claude Code. Multi-agent ingestion pipeline is in progress for GitHub Copilot, Gemini / Antigravity, OpenAI Codex, xAI Grok, Cursor, and Windsurf.

Multi-Agent Fleet

Live automated telemetry capture from autonomous terminal CLIs, IDE chat agents, and coding bots.

Token Trajectory

High-contrast multi-color tracking of prompt inputs, reasoning outputs, and prompt cache hits.

Session Explorer

Deep-dive into step durations, tool invocations, token costs, and full developer coding workflows.

Cost Governance

Real-time spend allocation by repository, model provider, and developer team with budget ceilings.

Quick Connect & Installation Guide

Read Full Docs

1. Install CLI

npm i -g tokentrail

or run via npx without install

2. Authenticate

tokentrail login

links terminal to workspace

3. Hook Claude Code

tokentrail hook --all

auto-attaches telemetry stream

Sub-millisecond latency • No code changes requiredClaude Code, Gemini, Copilot & SDK Setup →

Fleet Telemetry Availability:

Claude CodeLive & Publishing

GitHub CopilotIn Progress

Gemini / AntigravityIn Progress

OpenAI CodexIn Progress

xAI GrokIn Progress

Cursor / WindsurfIn Progress