{"id":47699422,"url":"https://github.com/orositororo/zero-employee-orchestrator","last_synced_at":"2026-04-07T06:01:28.295Z","repository":{"id":347214361,"uuid":"1171140922","full_name":"OrosiTororo/Zero-Employee-Orchestrator","owner":"OrosiTororo","description":"AI Orchestration Platform — Design · Execute · Verify · Improve","archived":false,"fork":false,"pushed_at":"2026-03-30T12:33:40.000Z","size":4593,"stargazers_count":2,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"master","last_synced_at":"2026-04-03T03:30:10.352Z","etag":null,"topics":["ai","fastapi","llm","multi-agent","orchestration","python","tauri","typescript"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/OrosiTororo.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":"CONTRIBUTING.md","funding":null,"license":"LICENSE","code_of_conduct":"CODE_OF_CONDUCT.md","threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":".github/SECURITY_SETUP_CHECKLIST.md","support":null,"governance":null,"roadmap":"ROADMAP.md","authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-03-02T23:08:44.000Z","updated_at":"2026-03-30T12:32:25.000Z","dependencies_parsed_at":"2026-04-02T18:02:36.402Z","dependency_job_id":null,"html_url":"https://github.com/OrosiTororo/Zero-Employee-Orchestrator","commit_stats":null,"previous_names":["orositororo/zero-employee-orchestrator"],"tags_count":2,"template":false,"template_full_name":null,"purl":"pkg:github/OrosiTororo/Zero-Employee-Orchestrator","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/OrosiTororo%2FZero-Employee-Orchestrator","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/OrosiTororo%2FZero-Employee-Orchestrator/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/OrosiTororo%2FZero-Employee-Orchestrator/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/OrosiTororo%2FZero-Employee-Orchestrator/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/OrosiTororo","download_url":"https://codeload.github.com/OrosiTororo/Zero-Employee-Orchestrator/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/OrosiTororo%2FZero-Employee-Orchestrator/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":31501903,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-04-07T03:10:19.677Z","status":"ssl_error","status_checked_at":"2026-04-07T03:10:13.982Z","response_time":105,"last_error":"SSL_connect returned=1 errno=0 peeraddr=140.82.121.5:443 state=error: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["ai","fastapi","llm","multi-agent","orchestration","python","tauri","typescript"],"created_at":"2026-04-02T17:02:33.313Z","updated_at":"2026-04-07T06:01:28.289Z","avatar_url":"https://github.com/OrosiTororo.png","language":"Python","funding_links":["https://github.com/sponsors/OrosiTororo"],"categories":[],"sub_categories":[],"readme":"**Language:** English | [日本語](docs/ja-JP/README.md) | [简体中文](docs/zh-CN/README.md) | [繁體中文](docs/zh-TW/README.md) | [한국어](docs/ko-KR/README.md) | [Português (Brasil)](docs/pt-BR/README.md) | [Türkçe](docs/tr/README.md)\n\n# Zero-Employee Orchestrator\n\n[![Stars](https://img.shields.io/github/stars/OrosiTororo/Zero-Employee-Orchestrator?style=flat)](https://github.com/OrosiTororo/Zero-Employee-Orchestrator/stargazers)\n[![Forks](https://img.shields.io/github/forks/OrosiTororo/Zero-Employee-Orchestrator?style=flat)](https://github.com/OrosiTororo/Zero-Employee-Orchestrator/network/members)\n[![Contributors](https://img.shields.io/github/contributors/OrosiTororo/Zero-Employee-Orchestrator?style=flat)](https://github.com/OrosiTororo/Zero-Employee-Orchestrator/graphs/contributors)\n[![License](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE)\n![Python](https://img.shields.io/badge/-Python-3776AB?logo=python\u0026logoColor=white)\n![FastAPI](https://img.shields.io/badge/-FastAPI-009688?logo=fastapi\u0026logoColor=white)\n![React](https://img.shields.io/badge/-React-61DAFB?logo=react\u0026logoColor=black)\n![TypeScript](https://img.shields.io/badge/-TypeScript-3178C6?logo=typescript\u0026logoColor=white)\n![Rust](https://img.shields.io/badge/-Rust-000000?logo=rust\u0026logoColor=white)\n![Docker](https://img.shields.io/badge/-Docker-2496ED?logo=docker\u0026logoColor=white)\n\n\u003e **The AI Meta-Orchestrator — Design · Execute · Verify · Improve**\n\n---\n\n**The orchestrator of orchestrators. One platform to command every AI framework, every LLM, every tool — under human approval and full audit.**\n\nZEO doesn't replace your AI tools — it unifies them. Connect CrewAI, AutoGen, LangChain, Dify, Claude Cowork, n8n, Zapier, and 34+ business apps under a single approval gate, audit trail, and security layer. Define workflows in natural language. Let multiple AI agents plan, execute, verify, and improve — all under your control.\n\nZEO itself is free and open source. LLM API costs are paid directly by users to each provider.\n\n---\n\n## Getting Started\n\n**Choose your path:**\n\n| Method | Best for | Time | API key needed? |\n|--------|----------|------|-----------------|\n| **[Desktop App](#-download-desktop-app)** | Non-technical users | 2 min | No (subscription mode) |\n| **[CLI (pip install)](#-quick-start-cli)** | Developers | 2 min | No (subscription or Ollama) |\n| **[Docker](#-docker)** | Self-hosting / production | 5 min | No (subscription or Ollama) |\n\n**System Requirements:** Python 3.12+ (CLI), Node.js 22+ (frontend dev), 4 GB RAM minimum. Ollama local models need 8 GB+ RAM.\n\n---\n\n## 🖥️ Download Desktop App\n\nPre-built desktop installers are available on the [Releases](https://github.com/OrosiTororo/Zero-Employee-Orchestrator/releases) page.\n\n| OS | File | Description |\n|---|---|---|\n| **Windows** | `-setup.exe` | Windows installer (x64) |\n| **macOS** | `.dmg` | macOS Universal (Intel + Apple Silicon) |\n| **Linux** | `.AppImage` | Portable (no install needed, amd64) |\n| **Linux** | `.deb` / `.rpm` | Debian/Ubuntu / Fedora/RHEL (amd64/x86_64) |\n\nAfter installation, a **setup wizard** will guide you through:\n1. **Language** — Choose English, 日本語, 中文, 한국어, Português, or Türkçe (changeable later in Settings)\n2. **LLM provider** — Pick how the AI runs (no API key needed for subscription mode)\n3. **First task** — Start using the platform immediately\n\n---\n\n## 🚀 Quick Start (CLI)\n\n### Step 1: Install\n\n```bash\n# PyPI (recommended)\npip install zero-employee-orchestrator\n\n# or from source\ngit clone https://github.com/OrosiTororo/Zero-Employee-Orchestrator.git\ncd Zero-Employee-Orchestrator \u0026\u0026 pip install .\n\n# or Docker (see Docker section below for details)\ndocker compose -f docker/docker-compose.yml up -d\n```\n\n### Step 2: Configure\n\nPick **one** of these options:\n\n```bash\n# Option A: No API key needed — uses free web AI services via g4f\nzero-employee config set DEFAULT_EXECUTION_MODE subscription\n\n# Option B: Fully offline — local models via Ollama (no internet needed)\nzero-employee config set DEFAULT_EXECUTION_MODE free\nzero-employee pull qwen3:8b\n\n# Option C: API key — best quality, pay-per-use to provider\nzero-employee config set OPENROUTER_API_KEY \u003cyour-key\u003e  # or GEMINI_API_KEY, etc.\n```\n\n\u003e **ZEO itself is free.** LLM costs (if any) are paid directly to each provider. See [USER_SETUP.md](USER_SETUP.md) for all options.\n\n### Step 3: Start\n\n```bash\n# Option A: start script (starts both backend + frontend automatically)\n./start.sh                       # macOS / Linux\n.\\start.ps1                      # Windows (PowerShell)\n# → Open http://localhost:5173\n\n# Option B: Manual start\nzero-employee serve              # Start the API server (port 18234)\ncd apps/desktop/ui \u0026\u0026 pnpm dev   # Start the frontend (port 5173) in another terminal\n# → Open http://localhost:5173\n\n# Option C: Chat mode only (no Web UI needed)\nzero-employee chat               # Default settings\nzero-employee local --model qwen3:8b  # Ollama\n```\n\n\u003e **Note:** `zero-employee serve` starts the API server only. The Web UI runs separately on port 5173. Use `start.sh` (or `start.ps1` on Windows) for the easiest setup.\n\n### Step 4: Verify\n\n```bash\nzero-employee health              # Check server status\nzero-employee models              # List available models\nzero-employee config list         # Review your settings\n```\n\n### Changing Language\n\nThe default language is English. Change it system-wide (CLI, AI responses, and Web UI all switch together):\n\n```bash\n# At startup\nzero-employee chat --lang ja      # Japanese\nzero-employee chat --lang zh      # Chinese\nzero-employee chat --lang ko      # Korean\nzero-employee chat --lang pt      # Portuguese\nzero-employee chat --lang tr      # Turkish\n\n# Persistently (saved to ~/.zero-employee/config.json)\nzero-employee config set LANGUAGE ja\n\n# At runtime (inside chat mode)\n/lang en                          # Switch to English\n/lang ja                          # Switch to Japanese\n/lang zh                          # Switch to Chinese\n/lang ko                          # Switch to Korean\n/lang pt                          # Switch to Portuguese\n/lang tr                          # Switch to Turkish\n```\n\nIn the desktop app, change language anytime via **Settings**.\n\n---\n\n## 🐳 Docker\n\n### API + Frontend (recommended)\n\n```bash\ndocker compose -f docker/docker-compose.yml up -d\n# → Open http://localhost:5173\n```\n\nThis starts three services: API server (port 18234), Frontend (port 5173), and a background worker.\n\n\u003e **Note:** Requires `SECRET_KEY` environment variable. Generate one: `python -c \"import secrets; print(secrets.token_urlsafe(32))\"`\n\n### API only\n\n```bash\ndocker compose up -d\n# → API available at http://localhost:18234/api/v1/\n```\n\nThis starts only the API server. Use this with the Desktop App or your own frontend.\n\n---\n\n## The Guides\n\n\u003ctable\u003e\n\u003ctr\u003e\n\u003ctd width=\"33%\"\u003e\n\u003ca href=\"docs/guides/quickstart-guide.md\"\u003e\n\u003cimg src=\"assets/images/guides/quickstart-guide.svg\" alt=\"Quickstart Guide\" /\u003e\n\u003c/a\u003e\n\u003c/td\u003e\n\u003ctd width=\"33%\"\u003e\n\u003ca href=\"docs/guides/architecture-guide.md\"\u003e\n\u003cimg src=\"assets/images/guides/architecture-guide.svg\" alt=\"Architecture Deep Dive\" /\u003e\n\u003c/a\u003e\n\u003c/td\u003e\n\u003ctd width=\"33%\"\u003e\n\u003ca href=\"docs/guides/security-guide.md\"\u003e\n\u003cimg src=\"assets/images/guides/security-guide.svg\" alt=\"Security Guide\" /\u003e\n\u003c/a\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"center\"\u003e\u003cb\u003eQuickstart Guide\u003c/b\u003e\u003cbr/\u003eFirst workflow, CLI basics.\u003c/td\u003e\n\u003ctd align=\"center\"\u003e\u003cb\u003eArchitecture Deep Dive\u003c/b\u003e\u003cbr/\u003e9-layer architecture, DAG, Judge Layer.\u003c/td\u003e\n\u003ctd align=\"center\"\u003e\u003cb\u003eSecurity Guide\u003c/b\u003e\u003cbr/\u003ePrompt defense, approval gates, sandbox.\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/table\u003e\n\n---\n\n## 📦 What's Inside\n\n```\nZero-Employee-Orchestrator/\n├── apps/\n│   ├── api/                  # FastAPI backend\n│   │   └── app/\n│   │       ├── core/               # Config, DB, security, i18n\n│   │       ├── api/routes/         # 46 REST API route modules\n│   │       ├── api/ws/             # WebSocket\n│   │       ├── models/             # SQLAlchemy ORM\n│   │       ├── schemas/            # Pydantic DTO\n│   │       ├── services/           # Business logic\n│   │       ├── repositories/       # DB I/O abstraction\n│   │       ├── orchestration/      # DAG, Judge, state machine\n│   │       ├── providers/          # LLM gateway, Ollama, RAG\n│   │       ├── security/           # IAM, secrets, sanitize, prompt defense\n│   │       ├── policies/           # Approval gates, autonomy boundaries\n│   │       ├── integrations/       # Sentry, MCP, external skills, Browser Assist\n│   │       └── tools/              # External tool connectors\n│   ├── desktop/              # Tauri v2 + React UI\n│   ├── edge/                 # Cloudflare Workers\n│   └── worker/               # Background workers\n├── skills/                   # 8 built-in skills\n├── plugins/                  # 16 plugin manifests (10 general + 6 role-based)\n├── extensions/               # 11 extension manifests\n│   └── browser-assist/\n│       └── chrome-extension/ # Chrome extension for Browser Assist\n├── packages/                 # Shared NPM packages\n├── docs/                     # Multi-language docs \u0026 guides\n│   ├── ja-JP/                # 日本語\n│   ├── zh-CN/                # 简体中文\n│   ├── zh-TW/                # 繁體中文\n│   ├── ko-KR/                # 한국어\n│   ├── pt-BR/                # Português (Brasil)\n│   ├── tr/                   # Türkçe\n│   └── guides/               # Architecture, security, quickstart guides\n└── assets/\n    └── images/\n        ├── guides/           # Guide header images\n        └── logo/             # Logo assets\n```\n\n---\n\n## 🏗️ 9-Layer Architecture\n\n```\n┌─────────────────────────────────────────┐\n│  1. User Layer       — Natural language input          │\n│  2. Design Interview — Requirements exploration        │\n│  3. Task Orchestrator — DAG decomposition \u0026 scheduling  │\n│  4. Skill Layer      — Specialized Skills + Context     │\n│  5. Judge Layer      — Two-stage + Cross-Model QA       │\n│  6. Re-Propose       — Rejection → dynamic DAG rebuild  │\n│  7. State \u0026 Memory   — Experience Memory               │\n│  8. Provider         — LLM Gateway (LiteLLM)           │\n│  9. Skill Registry   — Publish / Search / Import        │\n└─────────────────────────────────────────┘\n```\n\n---\n\n## 🎯 Key Features\n\n### Core Orchestration\n\n| Feature | Description |\n|---------|-------------|\n| **Design Interview** | Natural language requirements exploration and refinement |\n| **Spec / Plan / Tasks** | Structured intermediate artifacts — reusable, auditable, reversible |\n| **Task Orchestrator** | DAG-based planning with cost estimation and quality mode switching |\n| **Judge Layer** | Rule-based first pass + Cross-Model high-accuracy verification |\n| **Self-Healing / Re-Propose** | Automatic re-planning on failure with dynamic DAG reconstruction |\n| **Experience Memory** | Learns from past executions to improve future performance |\n\n### Meta-Orchestration (Orchestrator of Orchestrators)\n\n| Feature | Description |\n|---------|-------------|\n| **Agent Adapter** | Integrate CrewAI, AutoGen, LangChain, Dify agents as ZEO sub-workers — all under approval gates and audit |\n| **iPaaS Bridge** | Connect n8n / Zapier / Make workflows — ZEO becomes the judgment layer for existing automations |\n| **MCP Server** | Model Context Protocol — any MCP-compatible tool becomes a ZEO tool automatically |\n| **App Connector Hub** | 34+ apps (Obsidian, Notion, Google Workspace, Microsoft 365, Slack, etc.) — connectors for connectors |\n| **Tool-of-Tools** | Tools that connect to other tools (e.g., n8n connects to 400+ apps → ZEO connects to n8n → ZEO accesses all 400+ apps through one integration) |\n| **Multi-Model Freedom** | 26 model families via LiteLLM, Ollama local models, g4f web sessions — not locked to any single provider |\n\n### Extensibility\n\n| Feature | Description |\n|---------|-------------|\n| **Skill / Plugin / Extension** | 3-tier extensibility with full CRUD management |\n| **Role-Based Plugin Packs** | Pre-configured plugins per business role: Sales, Finance, HR, Legal, Marketing, Support |\n| **Natural Language Skill Generation** | Describe in plain language → AI auto-generates (with safety checks) |\n| **Skill Marketplace** | Community skill publishing, search, review, and installation |\n| **External Skill Import** | Import skills from GitHub repositories |\n| **Self-Improvement** | AI analyzes and improves its own skills (with approval) |\n| **Meta-Skills** | AI learns how to learn (Feeling / Seeing / Dreaming / Making / Learning) |\n\n### AI Capabilities\n\n| Feature | Description |\n|---------|-------------|\n| **Browser Assist** | Chrome extension overlay — AI sees your screen in real-time |\n| **Media Generation** | Image, video, audio, music, 3D — with dynamic provider registration |\n| **App Connector Hub** | 34+ apps (Obsidian, Notion, Google Workspace, Microsoft 365, etc.) |\n| **AI Tool Integration** | 55+ external tools across 21 categories |\n| **A2A Communication** | Peer-to-peer agent messaging, channels, and negotiation |\n| **Avatar AI** | Learns your decision patterns and evolves with you |\n| **Secretary AI** | Brain dump → structured tasks, bridges you and the AI org |\n| **Operator Profile** | Cowork-style about-me + global instructions — AI personalizes responses based on your role, priorities, and work style |\n| **Task Dispatch** | Cowork Dispatch-inspired background tasks — fire-and-forget with status polling |\n| **Repurpose Engine** | Auto-convert 1 content to 10 media formats |\n\n### Security\n\n| Feature | Description |\n|---------|-------------|\n| **Prompt Injection Defense** | 5 categories, 28+ detection patterns |\n| **Approval Gates** | 14 categories — including tiered browser automation (10 levels: navigate \u003c click \u003c type \u003c submit \u003c login \u003c payment) |\n| **File Sandbox** | AI can only access user-permitted folders (default: STRICT) |\n| **Data Protection** | Upload/download policy control (default: LOCKDOWN) |\n| **PII Protection** | Auto-detect and mask 13 categories of personal information |\n| **IAM** | Human/AI account separation, AI denied admin/secret access |\n| **Red-team Security** | 8-category, 20+ test self-vulnerability assessment |\n\n### Operations\n\n| Feature | Description |\n|---------|-------------|\n| **Multi-model Support** | Dynamic catalog, auto-fallback, per-task provider override |\n| **i18n** | 6 languages (EN / JA / ZH / KO / PT / TR) — UI, AI responses, CLI |\n| **Autonomous Operation** | Docker / Cloudflare Workers — runs when your PC is off |\n| **24/365 Scheduler** | 9 trigger types: cron, ticket creation, budget threshold, etc. |\n| **iPaaS Integration** | n8n / Zapier / Make webhook integration |\n| **Cloud Native** | AWS / GCP / Azure / Cloudflare abstraction layer |\n| **Governance \u0026 Compliance** | GDPR / HIPAA / SOC2 / ISO27001 / CCPA / APPI |\n\n---\n\n## 🔒 Security\n\nZEO is designed **security-first** with multi-layered defense:\n\n| Layer | Description |\n|-------|-------------|\n| **Prompt Injection Defense** | Detects and blocks instruction injection from external inputs (5 categories, 28+ patterns) |\n| **Approval Gates** | 14 categories of dangerous operations (send, delete, billing, permission changes) require human approval |\n| **Autonomy Boundaries** | Explicitly limits what AI can do autonomously |\n| **IAM \u0026 Tool Permissions** | Separate human/AI accounts; role-based tool permissions (5 default policies: secretary, researcher, reviewer, executor, admin) enforce least privilege per agent |\n| **Kill Switch** | Emergency halt of all active executions via UI button or API (`/kill-switch/activate`). Blocks new executions until resumed. |\n| **Tiered Judge** | Three-tier verification: LIGHTWEIGHT (rules only) → STANDARD (+policy) → HEAVY (+cross-model). Reduces cost for low-risk ops while maintaining full verification for high-risk ones. |\n| **Memory Trust** | Experience Memory entries track source type, trust level (0.0-1.0), verification status, and expiry. Only trustworthy memories (≥0.7, not expired) are used. |\n| **Secret Management** | Fernet encryption, auto-masking, rotation support |\n| **Sanitization** | Auto-removal of API keys, tokens, and PII |\n| **Security Headers** | CSP, HSTS, X-Frame-Options on all responses |\n| **Rate Limiting** | slowapi-based API rate limiting |\n| **Audit Logging** | All critical operations recorded (built-in from design, not bolted on) |\n\nSee [SECURITY.md](SECURITY.md) for vulnerability reporting.\n\n---\n\n## 🖥️ CLI Reference\n\n```bash\nzero-employee serve              # Start API server\nzero-employee serve --port 8000  # Custom port\nzero-employee serve --reload     # Hot reload\n\nzero-employee chat               # Chat mode (all providers)\nzero-employee chat --mode free   # Free mode (Ollama / g4f)\nzero-employee chat --lang en     # Language selection\n\nzero-employee local              # Local chat (Ollama)\nzero-employee local --model qwen3:8b --lang ja\n\nzero-employee models             # List installed models\nzero-employee pull qwen3:8b      # Download model\n\nzero-employee config list        # Show all settings\nzero-employee config set \u003cKEY\u003e   # Set a value\nzero-employee config get \u003cKEY\u003e   # Get a value\n\nzero-employee db upgrade         # Run DB migrations\nzero-employee health             # Health check\nzero-employee security status    # Security status\nzero-employee update             # Update to latest version\n```\n\n---\n\n## 🤖 Supported LLM Models\n\nManaged via `model_catalog.json` — swap models without code changes.\n\n| Mode | Description | Examples |\n|------|-------------|---------|\n| **Quality** | Highest quality | Claude Opus, GPT, Gemini Pro |\n| **Speed** | Fast response | Claude Haiku, GPT Mini, Gemini Flash |\n| **Cost** | Low cost | Haiku, Mini, Flash Lite, DeepSeek |\n| **Free** | Free | Gemini free tier, Ollama local |\n| **Subscription** | No API key needed | via g4f |\n\nPer-task provider override is supported — specify provider, model, and execution mode per task.\n\n---\n\n## 🧩 Skill / Plugin / Extension\n\n### 3-Tier Extensibility\n\n| Type | Description | Examples |\n|------|-------------|---------|\n| **Skill** | Single-purpose specialized processing | spec-writer, review-assistant, browser-assist |\n| **Plugin** | Bundles multiple Skills | ai-secretary, ai-self-improvement, youtube |\n| **Extension** | System integration \u0026 infrastructure | mcp, oauth, notifications, browser-assist |\n\n### Generate Skills with Natural Language\n\n```bash\nPOST /api/v1/registry/skills/generate\n{\n  \"description\": \"A skill that summarizes long documents into 3 key points\"\n}\n```\n\n18 dangerous patterns are auto-detected. Only skills passing safety checks are registered.\n\n---\n\n## 🌐 Browser Assist\n\nChrome extension overlay chat — AI sees your screen in real-time and guides you.\n\n- **Overlay Chat**: Chat UI directly on any website\n- **Real-time Screen Sharing**: AI sees what you see (no manual screenshots)\n- **Error Diagnosis**: AI reads error messages on screen and suggests fixes\n- **Form Assistance**: Step-by-step field-by-field guidance\n- **Privacy-first**: Screenshots processed temporarily, PII auto-masked, password fields blurred\n\n### Setup\n\n```\n1. Load extensions/browser-assist/chrome-extension/ in Chrome\n   → chrome://extensions → Developer mode → \"Load unpacked\"\n2. Click the chat icon on any website\n3. Ask questions or share your screen with the screenshot button\n```\n\n---\n\n## 🛠️ Tech Stack\n\n### Backend\n- Python 3.12+ / FastAPI / uvicorn\n- SQLAlchemy 2.x (async) + Alembic\n- SQLite (dev) / PostgreSQL (production)\n- LiteLLM Router SDK\n- bcrypt / Fernet encryption\n- slowapi rate limiting\n\n### Frontend\n- React 19 + TypeScript + Vite\n- shadcn/ui + Tailwind CSS\n- TanStack Query + Zustand\n\n### Desktop\n- Tauri v2 (Rust) + Python sidecar\n\n### Deploy\n- Docker + docker-compose\n- Cloudflare Workers (serverless)\n\n---\n\n## ❓ FAQ\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cb\u003eDo I need API keys to start?\u003c/b\u003e\u003c/summary\u003e\n\nNo. You can use subscription mode (no key needed) or Ollama for fully offline local AI. See the Quick Start section above.\n\u003c/details\u003e\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cb\u003eHow much does it cost?\u003c/b\u003e\u003c/summary\u003e\n\nZEO itself is free. LLM API costs are paid directly by you to each provider (OpenAI, Anthropic, Google, etc.). You can also run completely free with Ollama local models.\n\u003c/details\u003e\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cb\u003eCan I use multiple LLM providers simultaneously?\u003c/b\u003e\u003c/summary\u003e\n\nYes. ZEO supports per-task provider override — you can use Claude for high-quality spec reviews and GPT for fast task execution in the same workflow.\n\u003c/details\u003e\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cb\u003eIs my data safe?\u003c/b\u003e\u003c/summary\u003e\n\nZEO is self-hosted by design. Your data stays on your infrastructure. File sandbox defaults to STRICT, data transfer defaults to LOCKDOWN, and PII auto-detection is enabled by default.\n\u003c/details\u003e\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cb\u003eHow is this different from AutoGen / CrewAI / LangGraph?\u003c/b\u003e\u003c/summary\u003e\n\nZEO is a **business workflow platform**, not a developer framework. It provides human approval gates, audit logging, a 3-tier extensibility system, browser assist, media generation, and a complete REST API — all designed for running AI as an organization, not just chaining prompts.\n\u003c/details\u003e\n\n---\n\n## 🧪 Development\n\n```bash\n# Setup\ngit clone https://github.com/OrosiTororo/Zero-Employee-Orchestrator.git\ncd Zero-Employee-Orchestrator\npip install -e \".[dev]\"\n\n# Start (hot reload)\nzero-employee serve --reload\n\n# Test\npytest apps/api/app/tests/\n\n# Lint\nruff check apps/api/app/\nruff format apps/api/app/\n```\n\n---\n\n## 🤝 Contributing\n\nContributions are welcome.\n\n1. Fork → Branch → PR (standard flow)\n2. Security issues: follow [SECURITY.md](SECURITY.md) for private reporting\n3. Coding standards: ruff format, type hints required, async def\n\n---\n\n## 💜 Sponsors\n\nThis project is free and open source. Sponsors help keep it maintained and growing.\n\n[**Become a Sponsor**](https://github.com/sponsors/OrosiTororo)\n\n---\n\n## 🌟 Star History\n\n[![Star History Chart](https://api.star-history.com/svg?repos=OrosiTororo/Zero-Employee-Orchestrator\u0026type=Date)](https://star-history.com/#OrosiTororo/Zero-Employee-Orchestrator\u0026Date)\n\n---\n\n## 📄 License\n\nMIT — Use freely, modify as needed, contribute back if you can.\n\n---\n\n\u003cp align=\"center\"\u003e\n  \u003cstrong\u003eZero-Employee Orchestrator\u003c/strong\u003e — Run AI as an organization.\u003cbr\u003e\n  Built with security, auditability, and human oversight in mind.\n\u003c/p\u003e\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Forositororo%2Fzero-employee-orchestrator","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Forositororo%2Fzero-employee-orchestrator","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Forositororo%2Fzero-employee-orchestrator/lists"}