tinykit

Railway services are ephemeral. You must configure volumes to ensure your database and user projects are not lost on restart.

Required Volumes

1. Pocketbase Data

   • Mount Path: /app/pocketbase/pb_data
   • Purpose: Stores database, auth, and uploaded files.
   • Recommended Size: 1GB (Minimum 500MB; 5GB+ if handling many file uploads).

2. Workspace

   • Mount Path: /app/workspace
   • Purpose: Stores source code and compiled HTML output. Note: node_modules are NOT stored here.
   • Recommended Size: 1GB (Minimum 500MB; 2GB+ for heavy use).

Tinykit uses domain-based routing, allowing you to host multiple apps on a single server by pointing different domains to it. Each domain serves a specific app at the root path, while the Tinykit management tools are accessed via specific sub-paths.

Routing Example

• calculator.myserver.com/ → Serves the calculator app
• calculator.myserver.com/tinykit/studio → Opens the editor for the calculator app

URL Access Points

To run Tinykit directly with Node.js, clone the repository, install dependencies via npm, and run the development script. The dev server will automatically download PocketBase. After starting, visit http://localhost:5173/setup to complete the initial configuration.

# Clone and install
git clone https://github.com/tinykit-studio/tinykit.git
cd tinykit
npm install

# Start dev server (auto-downloads PocketBase)
npm run dev

The fastest way to deploy Tinykit is using the one-click Railway deployment. This requires no manual configuration. Once deployed, you can configure your LLM directly from the application interface or by adding your API key as an environment variable.

[![Deploy on Railway](https://railway.app/button.svg)](https://railway.com/deploy/tinykit?referralCode=RCPU7k&utm_medium=integration&utm_source=template&utm_campaign=generic)

To get a local instance of tinykit running quickly using Docker Compose, follow these steps:

1. Clone the repository and enter the directory.
2. Create your .env file from the example and add your LLM_API_KEY.
3. Navigate to the docker deployment directory and start the containers.

Once running, the App is available at http://localhost:3000 and the Builder at http://localhost:3000/tinykit.

# 1. Clone the repository
git clone https://github.com/tinykit-studio/tinykit.git
cd tinykit

# 2. Set environment variables
cp .env.example .env
# Edit .env and add your LLM_API_KEY

# 3. Run with Docker Compose
cd deploy/docker
docker-compose up -d

Docker is a universal option for deploying tinykit to any VPS (DigitalOcean, AWS, Azure, GCP, etc.). It provides full control and portability, with estimated costs between $6-15/mo depending on your VPS provider.

For detailed instructions, refer to the Docker Deployment Guide.

Deploy tinykit to Fly.io using the provided Dockerfile. You will need to install flyctl, launch the app, and set your LLM configuration as secrets.

# Install flyctl
curl -L https://fly.io/install.sh | sh

# Launch app
fly launch --dockerfile deploy/docker/Dockerfile

# Set secrets
fly secrets set LLM_API_KEY=sk-your-key
fly secrets set LLM_PROVIDER=openai
fly secrets set LLM_MODEL=gpt-4

# Deploy
fly deploy

If you prefer not to use Docker Compose, you can manually build and run the tinykit image. You must provide the required LLM environment variables and mount volumes for data persistence.

# Build image
docker build -f deploy/docker/Dockerfile -t crud-one .

# Run container
docker run -d \
  -p 3000:3000 \
  -e LLM_PROVIDER=openai \
  -e LLM_API_KEY=sk-your-key \
  -e LLM_MODEL=gpt-4 \
  -v crud_pb_data:/app/pocketbase/pb_data \
  -v crud_workspace:/app/workspace \
  --name crud-one \
  crud-one

Railway is the easiest deployment method, offering one-click deployment, automatic HTTPS, and built-in volumes. It is ideal for beginners who want managed infrastructure with a base cost of approximately $5/mo.

For detailed instructions, refer to the Railway Deployment Guide.

To run Tinykit locally with Docker, clone the repository, navigate to the docker deployment directory, and use docker-compose. This will start Tinykit on port 5173. After starting, visit http://localhost:5173/setup to complete the initial configuration.

git clone https://github.com/tinykit-studio/tinykit.git
cd tinykit/deploy/docker
docker-compose up -d