agenttrace AI coding agent observability dashboard

AI coding agent session history starts with local logs.

agenttrace reads the session files your coding agents already write and turns them into cost, token, time, and slow-run evidence without a hosted backend.

Install agenttrace View sample report View source

Claude Code Inspect local Claude Code sessions for cost, tokens, elapsed time, and slow spots.

Codex CLI Turn Codex CLI logs into reviewable session summaries and slow-run evidence.

Gemini CLI Compare Gemini CLI sessions across tokens, models, cost, duration, and tool latency.

More local logs Bring Qwen Code, Cline, Cursor exports, Aider, OpenCode, OpenClaw, Pi, Kimi CLI, and JSON/JSONL traces into the same local TUI.

What AI agent session history should answer.

Traditional logs show that something happened. Coding-agent session history should show what multiple agents spent across cost, tokens, and time, then explain why one run stalled or drifted into slow tool calls.

Cost and token history agenttrace --overview -f json

Slow-run report agenttrace --overview -f html -o agenttrace-overview.html

Health gate agenttrace --overview --fail-under-health 80 --fail-on-critical --max-tool-fail-rate 15

Local-first, source-aware, focused on spend and speed.

agenttrace is built for developers who want AI coding agent session history without sending prompts, code, or private logs to a hosted dashboard. It supports Claude Code, Codex CLI, Gemini CLI, Qwen Code, Cline, Aider, Cursor exports, Hermes Agent, OpenCode, OpenClaw, Pi, Oh My Pi, Kimi CLI, Copilot-style logs, and generic JSON/JSONL traces.