Agent Orchestrator

by Formatho

Desktop app for managing autonomous AI agents

A local-first Electron application with beautiful UI. Manage agents, schedule tasks, and automate workflows. No command line required. 100% private.

v0.1.0 Released
Electron Desktop App
100% Local
Open Source (MIT)
Privacy First
29,000+
Lines of Code
200+
Unit Tests
44
API Endpoints
15+
UI Components
3
Platforms
4+
LLM Providers

Desktop Application

Manage AI agents from a beautiful native desktop UI with premium light theme. No terminal needed - everything is visual and intuitive.

Agent Orchestrator
Navigation
Agents
TODOs
Schedule
Settings

Active Agents

daily-report
Generates sales report at 9am daily
Running
code-reviewer
Reviews PRs on GitHub
Idle
email-summarizer
Summarizes inbox every morning
Scheduled

Download

Available for macOS, Windows, and Linux. Self-contained applications with no external dependencies.

🍎

macOS (Intel)

x64 • DMG

~150MB
v0.1.0
Download
🍎

macOS (Apple Silicon)

arm64 • DMG

~150MB
v0.1.0
Download
🪟

Windows

x64 • EXE

~140MB
v0.1.0
Download
🐧

Linux

x64 • AppImage

~160MB
v0.1.0
Download

Features

Everything you need to manage autonomous AI agents. Open source, community-driven.

Desktop App (Electron)

Complete

Native desktop application with premium light theme UI. Manage agents visually - no command line required. Works on macOS, Windows, and Linux.

Agent Pool Management

Complete

Run multiple agents concurrently with resource limits. Create, pause, resume, and kill agents from dashboard. Full CRUD operations.

Persistent TODO Queue

Complete

Agents work through tasks autonomously. TODOs survive restarts with SQLite persistence. Track progress in real-time with priorities.

Cron Scheduling

Complete

Schedule agents to run at specific times with cron syntax. Daily reports, weekly summaries, monthly audits - fully automated.

Multi-LLM Support

Complete

Configure different LLMs for different agents. Supports OpenAI, Anthropic, LM Studio, Ollama, and custom providers. Per-agent model selection.

Skill System

Complete

Give agents real capabilities: file operations, web scraping, API calls, notifications. Fine-grained permissions system.

REST API + WebSocket

Complete

44 REST API endpoints for full control. WebSocket for real-time updates. SQLite database for persistence.

Optional Chat Interface

Planned

Talk to agents when needed. Provide input, debug behavior, or just check in. Chat is optional - agents work in background.

100% Local-First

Core

Your data stays on your machine. No cloud, no tracking, complete privacy. Works offline with local LLMs.

Built-in Skills

Give agents real capabilities. All skills run with permission controls.

file.read/write
Read and write local files
web.fetch/search
Scrape websites, search the web
shell.run
Execute shell commands (sandboxed)
http.get/post
Make HTTP requests to APIs
notify.desktop
Send desktop notifications
git.clone/push
Git operations
image.generate
Generate images (ComfyUI/DALL-E)
email.send
Send emails via SMTP

Modular Go Libraries

Each component is a standalone Go library. Use them independently or as a complete system. MIT licensed. 100% tested.

Ready to Automate?

Agent Orchestrator is released and ready to use. Star the repo, download the app, or contribute to the open source libraries.

Live User Count
1
Total Users this month
13,847