Skip to main content

A terminal UI tool for monitoring local AI coding agents (Claude Code, etc.)

Project description

Agentop

A terminal UI tool for monitoring local AI coding agents — like nvtop/htop, but for Claude/Codex.

Features

  • Real-time process monitoring (CPU, memory, uptime)
  • Claude Code usage + cost from local stats
  • Quota panels (beta) for Codex + Antigravity
  • Lightweight Textual TUI

Supported Agents

Agent Process Monitor Usage Stats Quota Status
Claude Code Stable
Antigravity Beta
OpenAI Codex Beta
OpenCode Beta

Supported Platforms (has been tested)

  • macOS
  • Linux

Installation (macOS / Linux)

PyPI

pip install agentop

From source

pip install git+https://github.com/dadwadw233/agentop.git

Quick Start

# TUI
agentop

# Or
python3 -m agentop

Detailed stats:

python3 show_stats.py

Data Sources

  • Claude stats: JSONL logs under ~/.config/claude/projects/ or ~/.claude/projects/ (override with CLAUDE_CONFIG_DIR, supports comma-separated paths)
  • Codex usage/quota: /usage API via Codex auth (~/.codex/auth.json)
  • Antigravity quota: Google Cloud Code API via Antigravity auth (local state db)
  • OpenCode stats: ~/.local/share/opencode/storage/ (message + session directories)

Changelog

0.2.0 (2026-01-25)

OpenCode Features:

  • Fixed time filtering bug - now shows all historical data (not just today)
  • Added time range support (Today/Week/Month/All) for all views
  • Implemented lazy loading - only computes aggregates needed for current view
  • Added index cache for faster incremental parsing
  • Performance optimizations for large datasets

TUI Improvements:

  • Redesigned overview dashboard with structured layout (Process Status, Session Stats, Token Usage)
  • Added progress bars for visual token usage comparison
  • Implemented pagination for handling large datasets
  • Dynamic page sizing based on available screen height
  • Enhanced table formatting with better alignment and smart truncation
  • Color gradients for usage intensity (cyan/magenta/blue)
  • Visual status indicators (🟢/⚪) with colored borders
  • Improved hint text with time range display and update timestamps
  • Consistent styling across all sub-views (Sessions, Projects, Models, Agents, Timeline)

Roadmap

  • More agents (TBD)
  • Config file (YAML)
  • History + export (CSV/JSON)
  • UI polish

Known Limitations

  • Claude stats can lag behind real time
  • Codex usage is fetched from the API (not local files)
  • Antigravity quota depends on account access
  • Antigravity refresh requires ANTIGRAVITY_OAUTH_CLIENT_SECRET or a fresh login
  • Proxy users: if you see “unknown scheme for proxy URL”, set AGENTOP_DISABLE_PROXY=1 or install httpx[socks]

License

MIT (see LICENSE)

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

agentop-0.2.1.tar.gz (44.4 kB view details)

Uploaded Source

Built Distribution

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

agentop-0.2.1-py3-none-any.whl (50.4 kB view details)

Uploaded Python 3

File details

Details for the file agentop-0.2.1.tar.gz.

File metadata

  • Download URL: agentop-0.2.1.tar.gz
  • Upload date:
  • Size: 44.4 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.9.6

File hashes

Hashes for agentop-0.2.1.tar.gz
Algorithm Hash digest
SHA256 1b3c08843940c36e6b8c4feb17dd1537aff0a8a8242b687c4ef81c2bfb6195a5
MD5 878d9052fbf84421f91fa9d429d4795a
BLAKE2b-256 b028c0dc7c39c64c4dbdbc319b40168000a7c08fa60dfc460c7af7944ef5804c

See more details on using hashes here.

File details

Details for the file agentop-0.2.1-py3-none-any.whl.

File metadata

  • Download URL: agentop-0.2.1-py3-none-any.whl
  • Upload date:
  • Size: 50.4 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.9.6

File hashes

Hashes for agentop-0.2.1-py3-none-any.whl
Algorithm Hash digest
SHA256 bdaefabb3cbb5b1730ac0b5bca287dccab26cac9132506ce9d161f615f716515
MD5 b4feba589b8ea2655d237cc93003a155
BLAKE2b-256 a9cd27f57420e9681906273babd83fb02b1e70460fc9a3be78d0b6da6643cf60

See more details on using hashes here.

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