🚀 AutoFlow — AI-Powered Workflow Automation

Describe what you want in plain English. AutoFlow's AI builds and runs the workflow.

AutoFlow is an open-source, AI-first workflow automation platform. Instead of manually connecting nodes and configuring webhooks, you describe your automation in natural language — AutoFlow's AI agent interprets your intent, builds the workflow visually, and executes it on a schedule or trigger.

Think Zapier or n8n, but you just type what you want and AI builds it.

✨ Features

• 🤖 AI Workflow Builder — Describe your workflow in plain English; AI generates the full pipeline with smart configs, filters, and template variables
• 🧠 Multi-Model AI — Choose from 6+ AI models: Ollama Cloud (gpt-oss:120b, deepseek-v3.1:671b, qwen3-coder:480b, glm-5.1, glm-5.2), OpenAI, or Anthropic
• 🔗 20+ Integrations — HTTP, Email, Slack, Discord, Telegram, Database, RSS, HackerNews, Weather, GitHub, X/Twitter, Webhook, Google Sheets, AI Sentiment, and more
• 📅 Scheduler — Cron-based triggers, event-based triggers, webhook triggers
• 📊 Visual Editor — ReactFlow-based drag-and-drop workflow editor with live execution
• 🔐 Self-Hosted — Your data stays on your server. Full privacy.
• 🏗️ Extensible — Plugin system for custom nodes and integrations
• 📈 Analytics — Execution history, success rates, performance metrics
• 🎯 12 Templates — Pre-built workflows for weather, news, social, AI, DevOps, and data use cases
• 🌐 REST API — Full programmatic access for developers
• 🐳 Docker Ready — One command deployment with docker-compose

🧠 AI Models

AutoFlow supports multiple AI providers for workflow generation and AI nodes:

Provider	Models	Type
Ollama Cloud	gpt-oss:120b, deepseek-v3.1:671b, qwen3-coder:480b, glm-5.1, glm-5.2, lfm2.5:8b	Cloud + Local
OpenAI	gpt-4o-mini, gpt-4o	Cloud
Anthropic	Claude Sonnet 4	Cloud
Rule-Based	Built-in pattern matcher	Offline fallback

The AI Builder generates workflows with:

• Smart node configurations (real URLs, cron expressions, API endpoints)
• Template variables ({{weather.temperature}}, {{data.title}})
• Conditional filters and branching
• Multi-step pipelines with AI processing

🏗️ Architecture

┌─────────────────────────────────────────────────────┐
│                   React Frontend                     │
│  ┌───────────┐ ┌───────────┐ ┌──────────────────┐  │
│  │ Dashboard  │ │ Workflow  │ │ AI Builder +     │  │
│  │ & Analytics│ │ Editor    │ │ Model Selector   │  │
│  └───────────┘ └───────────┘ └──────────────────┘  │
└──────────────────────┬──────────────────────────────┘
                       │ REST API
┌──────────────────────▼──────────────────────────────┐
│                  FastAPI Backend                     │
│  ┌──────────┐ ┌──────────┐ ┌──────────────────────┐ │
│  │ Workflow │ │ AI Agent │ │  Integration Layer   │ │
│  │ Engine   │ │ (Ollama) │ │  (20+ connectors)    │ │
│  └──────────┘ └──────────┘ └──────────────────────┘ │
│  ┌──────────┐ ┌──────────┐ ┌──────────────────────┐ │
│  │Scheduler │ │  Auth &  │ │  Templates & Seeds   │ │
│  │ (Cron)   │ │  Users   │ │  (12 pre-built)     │ │
│  └──────────┘ └──────────┘ └──────────────────────┘ │
└──────────────────────┬──────────────────────────────┘
                       │
              ┌────────▼────────┐
              │  SQLite / PostgreSQL  │
              └────────────────┘

🚀 Quick Start

Docker (Recommended)

# Clone
git clone https://github.com/pitee33/autoflow.git
cd autoflow

# Set up your AI keys
cp .env.example .env
# Edit .env — add your Ollama API key (https://ollama.com/settings)

# Launch
docker-compose up -d

Visit http://localhost:3000 — register and start building workflows!

Manual Setup

# Backend
cd backend
python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env  # Add your API keys
uvicorn app.main:app --reload --port 8000

# Frontend (new terminal)
cd frontend
npm install
npm run dev

Visit http://localhost:5173

Environment Variables

Variable	Description	Default
OLLAMA_API_KEY	Ollama Cloud API key	—
OLLAMA_BASE_URL	Ollama API endpoint	https://ollama.com/v1
OLLAMA_MODEL	Default AI model	gpt-oss:120b-cloud
OPENAI_API_KEY	OpenAI API key (optional)	—
ANTHROPIC_API_KEY	Anthropic API key (optional)	—
TELEGRAM_BOT_TOKEN	Telegram bot token for Telegram nodes	—
TELEGRAM_DEFAULT_CHAT_ID	Default Telegram chat ID	—
SECRET_KEY	JWT secret key	change-this

💰 Monetization (Cloud SaaS)

AutoFlow is open-source and free to self-host. The hosted AutoFlow Cloud offers:

• Free — $0: 3 workflows, 100 runs/mo, rule-based AI
• Starter — $19/mo: 25 workflows, 5K runs/mo, Ollama Cloud AI
• Pro — $49/mo: Unlimited workflows, 50K runs/mo, all AI models
• Enterprise — $199/mo: On-prem support, SLA, SSO, custom integrations

🛣️ Roadmap

• Core workflow engine with topological sort
• AI workflow builder (Ollama Cloud + OpenAI + Anthropic)
• Visual workflow editor (ReactFlow)
• 20+ integrations (HTTP, Telegram, Slack, Discord, Weather, RSS, etc.)
• 12 pre-built templates
• Docker Compose deployment
• Model selector (6 Ollama models)
• Telegram bot integration with template variables
• AutoFlow Cloud (hosted SaaS)
• Marketplace for workflow templates
• Team collaboration features
• Mobile app
• 100+ integrations

📄 License

MIT — do whatever you want. Fork it, sell hosted versions, build a business on it.

🤝 Contributing

PRs welcome! See CONTRIBUTING.md.

---

⭐ If this project helps you, give it a star!

GitHub: https://github.com/pitee33/autoflow