fagoon-upgrade 2.0.8

Fagoon AI Agents Workflow

Fagoon Upgrade

Self-hosted AI platform with workflow automation, agents, chat, and API generation.

Build AI workflows visually, publish them as REST APIs, deploy agents, and chat with LLMs — all running on your own machine.

---

Quick Start

Prerequisites

• Docker Desktop — Download here
• Python 3.11+ — Download here
• pipx (recommended) — Install with: pip install pipx

Install & Run

# Install the CLI
pipx install fagoon-upgrade

# Start the platform
fagoon up

On first run, you'll see a setup wizard:

  Welcome to Fagoon!
  Let's set up your AI platform.

  How would you like to power your AI?
  1) I have an API key (OpenAI, Gemini, Groq, etc.)
  2) Use local Ollama model (free, runs on your machine)
  3) Skip for now (configure later in the UI)

  Choose [1/2/3]:

Access the Platform

Service	URL
Frontend	http://localhost:3000
Backend	http://localhost:8000

---

Features

• Visual Workflow Builder - Drag-and-drop canvas with 20+ node types
• Workflow-as-API - Publish any workflow as a REST endpoint with API key auth
• AI Agents - Conversational agents with memory and tool use
• Chat Interface - Chat with any LLM provider through a unified UI
• Multi-Provider - OpenAI, Gemini, Groq, Anthropic, Ollama
• Self-Hosted - Your data stays on your machine

---

CLI Commands

fagoon up                    # Start the platform (lite mode)
fagoon up --full             # Start with Redis + Celery (multi-worker)
fagoon up --ollama           # Start with local Ollama LLM
fagoon down                  # Stop the platform
fagoon logs -f               # Follow live logs
fagoon status                # Check if running
fagoon update                # Pull latest version and restart
fagoon config set KEY=VALUE  # Set configuration
fagoon config show           # Show current config
fagoon db set-url URL        # Switch database

---

Workflow-as-API

Publish any workflow as a callable REST API:

curl -X POST http://localhost:8000/api/v1/workflow-api/your-slug/execute \
  -H "X-API-Key: wfapi_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{"input": "Hello from my app"}'

Response:

{
  "success": true,
  "execution_id": "uuid",
  "status": "COMPLETED",
  "output": {
    "output_text": "AI generated response",
    "model": "gemini-2.5-flash"
  },
  "usage": { "duration_ms": 2500 }
}

---

Deployment Modes

Lite Mode (Default)

Single-user, no Redis/Celery. Perfect for personal use.

fagoon up

Full Mode

Multi-worker with Redis + Celery. For production.

fagoon up --full

With Ollama (Local LLM)

Free, offline AI using local models.

fagoon up --ollama

---

Docker Compose (Direct)

services:
  app:
    image: ghcr.io/fagoon-ai/upgrade:2.0.4
    ports:
      - "8000:8000"
      - "3000:3000"
    environment:
      LITE_MODE: "true"
      DATABASE_URL: postgresql+asyncpg://fagoon:fagoon@db:5432/fagoon
    depends_on:
      db: { condition: service_healthy }

  db:
    image: pgvector/pgvector:pg16
    environment:
      POSTGRES_USER: fagoon
      POSTGRES_PASSWORD: fagoon
      POSTGRES_DB: fagoon
    volumes:
      - fagoon_db:/var/lib/postgresql/data
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U fagoon"]
      interval: 5s
      timeout: 5s
      retries: 10

volumes:
  fagoon_db:

---

Environment Variables

Variable	Default	Description
LITE_MODE	true	true = single-process, false = Redis+Celery
DATABASE_URL	auto	PostgreSQL connection string
JWT_SECRET	auto	Secret for JWT tokens
GEMINI_API_KEY	—	Google Gemini API key
OPENAI_API_KEY	—	OpenAI API key
GROQ_API_KEY	—	Groq API key
OLLAMA_BASE_URL	—	Ollama server URL
WEB_CONCURRENCY	1	Number of workers

---

Troubleshooting

"Cannot connect to Docker daemon" — Start Docker Desktop and wait for it to load.

"Port 3000/8000 already in use" — Stop other services on those ports.

Platform won't start:

fagoon logs -f     # Check logs
fagoon down        # Stop
fagoon up          # Restart

Reset everything:

fagoon down
docker volume rm fagoon_db fagoon_data
fagoon up

---

Tech Stack

• Backend: Python, FastAPI, SQLAlchemy, PostgreSQL, pgvector
• Frontend: Next.js, React, TailwindCSS, React Flow
• Workflow Engine: Custom DAG executor with 20+ node types
• Task Queue: Celery + Redis (full mode)
• Container: Docker, Docker Compose

---

Links

• Website: fagoonai.com
• PyPI: pypi.org/project/fagoon-upgrade
• Docker: docker pull ghcr.io/fagoon-ai/upgrade:2.0.4
• GitHub: github.com/Fagoon-AI/upgrade

License

MIT