VasiHemanth/tokentelemetry

Token telemetry dashboard for AI autonomous and coding agents — tracks tokens, sessions, tool calls & reasoning across Hermes agent, Claude Code, Antigravity CLI, Codex & more. 100% local.

What it solves

TokenTelemetry provides a local, zero-config observability dashboard for AI coding agents and autonomous agents. It solves the problem of "token blindness"—where users cannot easily track the cost, token usage, and internal reasoning steps of various AI agents (like Claude Code, Cursor, and Gemini CLI) across different projects and models in a single, unified view.

How it works

Instead of requiring SDKs or cloud accounts, TokenTelemetry reads the local session logs that AI agents already write to the machine. It parses these logs in real-time to extract token counts, tool calls, and reasoning chains, presenting them through a FastAPI backend and a Next.js frontend. It is designed to be 100% local, ensuring that prompts and usage data never leave the user's machine.

Who it’s for

Developers and power users who use multiple AI coding agents or autonomous agents (such as the Hermes Agent) and want to monitor their API spend, track token efficiency, and audit the internal logic of their AI-driven workflows without sending their data to a third-party cloud.

Highlights

  • Multi-Agent Support: Unified tracking for a wide array of agents including Claude Code, Cursor, Copilot, Gemini CLI, and Hermes Agent.
  • Local-First Privacy: All logs, sessions, and costs are processed locally; no cloud account or signup is required.
  • Detailed Tracing: Waterfall views of system prompts, reasoning steps, and tool calls to see exactly how an agent arrived at a result.
  • Cost & Budget Management: Real-time cost tracking with observational budgets and alerts to monitor spend per project or agent.
  • Plan Limit Tracking: Reads actual provider quota windows (e.g., session or weekly limits) directly from the machine's login state.
  • Hermes Agent Integration: A dedicated dashboard for autonomous agents featuring source platform tracking, gateway health, and subagent cards.

Related

  • Project
    xiufengsun/TokenTracker

    A local-first token usage and cost dashboard for 41 AI coding tools that tracks token burn and costs without requiring API keys or cloud accounts.

  • Project
    soulduse/ai-token-monitor

    A system tray application for macOS and Windows that tracks real-time token usage and costs for AI coding tools by parsing local session logs.

  • Project
    vivekchand/clawmetry

    ClawMetry is a local, zero‑config dashboard that reads the session logs of AI coding agents (supports 33 runtimes) and visualises timelines, tool calls, token usage, costs and health alerts. It runs as a Python package (`pip install clawmetry && clawmetry`), opens a web UI on localhost, and keeps all session data on the machine unless you explicitly upload an encrypted snapshot. Free for four runtimes; paid plans unlock the rest, cloud sync, approvals, and advanced anomaly detection.

  • Project
    mikehasa/agentacct

    A local-first agent work ledger that tracks token usage, cost estimates, and activity timelines for AI coding agents using local session logs.

  • Project
    f/agentlytics

    A local analytics dashboard that unifies conversation history, token costs, and usage patterns from 17 different AI coding editors into one place.