lazyagents

<p align="center"><img src="docs/assets/logo.svg" width="160" alt="lazyagents logo: a sloth asleep in a hammock while three robots carry mini terminals"></p> <h1 align="center">lazyagents</h1> <p align="center"><strong>The lazygit for AI coding agents.</strong><br>Manage your coding agents from one terminal.</p> <p align="center">Claude Code · Codex · Gemini CLI · OpenCode · Pi · Crush · Hermes Agent</p> <p align="center"> <a href="https://github.com/rogeriojunior31/lazyagents/actions/workflows/ci.yml"><img src="https://github.com/rogeriojunior31/lazyagents/actions/workflows/ci.yml/badge.svg" alt="CI"></a> <a href="https://github.com/rogeriojunior31/lazyagents/releases"><img src="https://img.shields.io/github/v/release/rogeriojunior31/lazyagents" alt="Latest release"></a> <a href="https://rogeriojunior31.github.io/en/docs/lazyagents/"><img src="https://img.shields.io/badge/docs-read%20online-f2984a" alt="Documentation"></a> </p> <p align="center"><b>English</b> · <a href="README.pt-br.md">Português</a></p> <!-- check-english:allow (language name) -->

Different agents keep skills, conversations and configuration in different places. lazyagents brings that work into one library and a browsable history while keeping each tool's native formats.

Features · Installation · Quick start · Support · Documentation

<p align="center"><a href="docs/assets/demos/hero.mp4"><img src="docs/assets/demos/hero.gif" width="1200" alt="Overview with sample data: skills, transcripts and usage"></a></p>

Recorded with sample data at 190×50 cells. Click to open the original-resolution video. This checkout includes Unreleased changes; the website and install command below follow published release 0.2.3.

Features

| What you can do | Guide | Demo | |---|---|---| | Keep one skill library, enable per agent and review batch actions | Skills | Video | | Search, resume, read and export conversations; open verified child sessions | Sessions | Video | | Save API/local-model connections and review changes before applying | Providers | Video | | Manage hooks and review commands installed in each agent | Hooks | Video | | Inspect backups, restore configuration and review previous operations | Recovery | Video | | Track limits, tokens and costs with explicit sources and coverage | Usage | Video | | Check detection, capabilities and local problems with doctor | Agents | Video |

<details> <summary>Watch the module demos at full content width</summary>

Skills

Enable skills per agent and review batch destinations.

Guide · Open GIF

Sessions

Search and read transcripts, including verified child sessions.

Guide · Open GIF

Providers

Review Gemini’s two-file changes and connection templates.

Guide · Open GIF

Hooks

Review installation before enabling a command.

Guide · Open GIF

Recovery

Inspect backups and retained operation history.

Guide · Open GIF

Usage

Compare agents, projects and models with cost evidence.

Guide · Open GIF

Agents

Inspect detection and run local diagnostics.

Guide · Open GIF

</details>

Installation

With Go 1.26 or newer:

go install github.com/rogeriojunior31/lazyagents@v0.2.3

Or download a Linux, macOS or Windows binary from Releases and check SHA256SUMS. The installation and upgrade guide covers platforms, optional tools and local Homebrew/Arch recipes. The public tap and AUR package have not been published yet.

Use the explicit version above: during renumbering, @latest may resolve a retired higher version. lazyagents manages existing agents; install their CLIs separately.

Quick start

lazyagents --version
lazyagents doctor --local   # check detection and files without external queries
lazyagents                  # open the interface

1. In Skills, import a folder or repository with i and review activation destinations.
2. In Sessions, select a conversation and press v to read it or enter to resume in the agent's CLI.
3. Use : for commands such as providers, recovery and agents doctor.

| Key | Action | |---|---| | tab / shift+tab | switch tabs | | ↑ / ↓ | select | | ? | show help for the current screen | | : | open the command palette | | esc | go back or cancel | | q | quit from the main screen |

The key reference covers actions by context. The CLI provides headless listings and operations, including JSON output.

Support

Claude Code, Codex, Gemini CLI, OpenCode, Pi, Crush and Hermes Agent have different capabilities. Check the generated matrix and verified versions before relying on a specific feature. Hermes provides skills; Claude Desktop appears in detection only.

Skill activation uses symlinks; Windows requires the corresponding permission. Agent configuration writes have backups and review; previews hide secrets. Subscription limits can query the network when requested. See Safety and data and Recovery.

The project is pre-1.0. Commands and formats may change between minor versions; see the CHANGELOG.

Documentation

Read online · This revision's index · Portuguese

| You want to… | Reference | |---|---| | Install, upgrade and get started | Getting started | | Follow a module workflow | Task guides | | Find a command, key or agent capability | CLI · Keys · Agents | | Configure paths, tabs and appearance | Configuration · Themes | | Understand writes, backups or a problem | Safety · Troubleshooting | | Add external tabs and commands | Plugins |

The README introduces the product; guides explain workflows and limits. Command and capability references are generated from code. Tests check links and reviewed translations; the website publishes release documentation.

Contributing

See CONTRIBUTING and the architecture. BACKLOG holds the consolidated roadmap, delivered work, next tasks and acceptance criteria. Security issues follow SECURITY.

License

MIT. The themes come from SP Night and community palettes, under their own licenses (notices).