Skip to main content

Agent Eye dashboard for monitoring GitHub Copilot CLI sessions in real-time

Project description

Agent Eye

GitHub GitHub release

CI Release

Publish to PyPI PyPI version OpenAPI

A local web dashboard that monitors all your GitHub Copilot CLI and Claude Code sessions in real-time. Designed for power users running multiple AI coding sessions simultaneously.

[!TIP] The dashboard works out of the box by reading events.jsonl files from your Copilot session directories. For richer session history (summaries, checkpoints), enable the SESSION_STORE experimental feature: add "experimental": true to ~/.copilot/config.json and start a new Copilot session.

Dashboard Screenshot

Installation

Option 1: From PyPI

pip install agenteye-app

Option 2: From Source

# Clone the repo
git clone https://github.com/JeffSteinbok/agenteye.git
cd agenteye

# Install in editable mode
pip install -e .

Usage

Native App (Recommended)

Run as a native desktop app with system tray integration:

# Start the tray app (window + tray icon)
agenteye app

# Start minimized to tray
agenteye app --hidden

# Custom port
agenteye app --port 8080

The tray app provides:

  • Native window with dark/light title bar matching app theme
  • System tray icon with quick access menu
  • Close (X) minimizes to tray, quit from tray menu exits
  • Native Windows notifications (no browser permission needed)

Browser Mode

Run as a background server and open in your browser:

# Start the dashboard
agenteye start

# Start in background
agenteye start --background

# Check status
agenteye status

# Stop
agenteye stop

# Upgrade to the latest version (restarts automatically if running)
agenteye upgrade

Open http://localhost:5111 in your browser.

Autostart at Login (Windows)

# Start tray app on login (recommended)
agenteye autostart --mode app

# Start background server on login
agenteye autostart --mode server

# Custom port
agenteye autostart --mode app --port 8080

# Remove login startup
agenteye autostart-remove

Features

Session States

  • Working / Thinking (green) — session is actively running tools or reasoning
  • Waiting (yellow) — session needs your input (ask_user or ask_permission pending)
  • Idle (blue) — session is done and ready for your next task

Key Features

  • Native tray appagenteye app runs as a native desktop application with system tray integration, eliminating the need for a separate browser tab
  • Dark/light title bar — window title bar automatically matches your chosen theme
  • Native notifications — Windows toast notifications with proper app name and icon (no browser permission prompts)
  • Start hidden--hidden flag starts the app minimized to tray (great for autostart)
  • Claude Code support — automatically discovers Claude Code sessions from ~/.claude/projects/. Active Claude sessions appear alongside Copilot sessions with a ✦ Claude badge.
  • Cross-machine sync — see active sessions from all your machines in one dashboard, powered by OneDrive or any cloud-synced folder. See Cross-Machine Sync for details.
  • Settings menu — ☰ hamburger menu in the header with toggles for autostart-on-login and remote sync.
  • Upgrade commandagenteye upgrade stops the server, upgrades via pip, and restarts automatically.
  • Desktop notifications — get alerts when sessions transition between states
  • Focus window — bring an active session's terminal to the foreground with one click
  • Restart commands — copy-pasteable copilot --resume <id> commands for every session
  • Waiting context — shows what a waiting session is asking (e.g. the ask_user question and choices)
  • Background tasks — shows count of running subagents per session
  • Session details — click any session to see checkpoints, recent tool output, references, and conversation history
  • Tile & List views — compact card grid or detailed expandable rows
  • 9 color palettes and light/dark mode

Cross-Machine Sync

See active sessions from all your machines in one dashboard — powered by OneDrive, Google Drive, or any cloud-synced folder. No Git commits needed.

How it works:

  • On each poll cycle, the dashboard exports your active sessions as JSON files to a shared cloud folder
  • Other machines read those files and display them in a "Remote Sessions" section under Active
  • Each machine only writes to its own subfolder — no sync conflicts

Auto-detection (priority order):

  1. OneDriveCommercial (preferred — prevents data leakage to personal accounts)
  2. OneDriveConsumer
  3. User Documents folder

Configuration (~/.copilot/dashboard-config.json):

{
  "sync": {
    "enabled": true,
    "folder": "D:\\MyCloudSync"
  }
}
  • Set "enabled": false to disable sync entirely
  • Set "folder" to override auto-detection with a specific path

What remote sessions show:

  • Live state indicators (working, waiting, idle)
  • Session summary, intent, branch, MCP servers, turn/checkpoint counts
  • Machine name badge (e.g. 🖥️ LAPTOP-HOME)

What remote sessions don't show:

  • No detail drill-down (checkpoints, turns, files)
  • No focus or kill actions (those are local-only)
  • No past/previous sessions from remote machines

Prerequisites

Package Purpose
fastapi Web framework with auto-generated OpenAPI docs
uvicorn ASGI server
pywinauto Window focus and process detection (Windows-only)
pywebview Native window for tray app
pystray System tray icon
plyer Native OS notifications

All are installed automatically via pip install agenteye-app.

For more details on architecture, data sources, and API endpoints, see DEVELOPMENT.md.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

agenteye_app-1.0.1.tar.gz (1.4 MB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

agenteye_app-1.0.1-py3-none-any.whl (1.4 MB view details)

Uploaded Python 3

File details

Details for the file agenteye_app-1.0.1.tar.gz.

File metadata

  • Download URL: agenteye_app-1.0.1.tar.gz
  • Upload date:
  • Size: 1.4 MB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.12

File hashes

Hashes for agenteye_app-1.0.1.tar.gz
Algorithm Hash digest
SHA256 729afd6b119bbe40ecdb48beda20b3a0a51da3d390f46d60f5175df9bd624e45
MD5 c97c7354b26053de365f422f515386f0
BLAKE2b-256 41adb2a279dd8bfe5c374840ab304fb8a06d244205ef61e99a48e666417a7831

See more details on using hashes here.

Provenance

The following attestation bundles were made for agenteye_app-1.0.1.tar.gz:

Publisher: publish-pypi.yml on JeffSteinbok/agenteye

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file agenteye_app-1.0.1-py3-none-any.whl.

File metadata

  • Download URL: agenteye_app-1.0.1-py3-none-any.whl
  • Upload date:
  • Size: 1.4 MB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.12

File hashes

Hashes for agenteye_app-1.0.1-py3-none-any.whl
Algorithm Hash digest
SHA256 277039c7a663f3d0c54c1588d1bc9f211aa0998c50ae0e99b887e175a258ba94
MD5 026bdcd8af15602034041f1a1190bfc8
BLAKE2b-256 e5dfc643710b9e8f62ad437ca4c455710d21348e545133cd435f049465de8248

See more details on using hashes here.

Provenance

The following attestation bundles were made for agenteye_app-1.0.1-py3-none-any.whl:

Publisher: publish-pypi.yml on JeffSteinbok/agenteye

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page