Step 1 Connect your agent
Add LogTrace as a local MCP server for tools such as Codex, Claude Code, Cursor, and other compatible clients.
AI agent audit monitor
Local-first audit trails for AI coding agents.
LogTrace records AI coding agent activity on your computer, including shell commands, file reads, agent notes, output, and risk labels. Review what happened after a session while keeping the audit trail in local storage.
LogTrace
| 3 entries Risk Level All Action All Project my-website Search...
| Time | Action | Description | Risk |
|---|---|---|---|
| --:--:-- | Shell Execute | Admin-level command | CRITICAL |
| --:--:-- | Read File | Secret-like file access | WARNING |
| --:--:-- | Note | Agent note recorded | INFO |
How it works
Use LogTrace alongside MCP-compatible coding agents to capture the actions that matter during AI-assisted development.
Step 2 Record activity locally
Save the captured shell commands, file contents, agent notes, and output data to your computer.
Step 3 Review risky actions
Find entries flagged for destructive commands, secret reads, package installs, downloads, or access outside the project.
Built for local agent review
LogTrace focuses on the core review loop developers need after an AI agent works in a project.
Organized by project
Save captured agent activity with project context, so it will be easier to find the logs you need later.
Explainable risk labels
Review why an entry was flagged, with the relevant command, file access, output, or project boundary shown close to the label.
Private by default
Logs stay on your computer in local storage. No hosted dashboard or cloud sync is required.
Optional AI review
When needed, run AI analysis of activity history with your own AI provider key and review estimated API usage.
Use Cases
LogTrace helps developers review AI coding agent activity when they need a local record of what changed, what ran, and what deserves attention.
Review agent sessions
Look back through shell commands, file reads, notes, and output after an AI-assisted coding session.
Check risky actions
Find destructive commands, secret-like file access, package installs, downloads, or project boundary access before you trust the result.
Keep local audit records
Keep project-specific activity history recorded during development on your computer without relying on a hosted dashboard.
Review agent activity.
Download LogTrace for Windows, or review pricing before adding it to your development workflow.