Usage

This guide shows you how to set up and use Brizy AI with Docker. Brizy AI runs as a containerized service and requires specific environment variables to function properly.

Quick Setup

Option 1: Docker Run Command

Run Brizy AI directly using Docker with all required environment variables:

# Only for development
docker run -d \
  --name brizy-ai \
  -p 80:80 \
  -p APP_HTTP_SCHEME=http \
  -p APP_ENV=dev \
  -e APP_URL=http://localhost \
  -e APP_ENCODE_KEY=base64:dGVzdA== \
  -e AMAZON_S3_REGION=Bucket Region \
  -e AMAZON_S3_BUCKET=Bucket Name \
  -e AMAZON_S3_KEY=EXAMPLE KEY \
  -e AMAZON_S3_SECRET=EXAMPLE KEY \
  -e APP_GPT_MODEL_HEADING=EXAMPLE KEY \
  -e APP_GPT_MODEL_HEADING_LOCATION=EXAMPLE KEY \
  -e APP_GPT_MODEL_SUBHEAD=EXAMPLE KEY \
  -e APP_GPT_MODEL_SUBHEAD_HEADING=EXAMPLE KEY \
  -e APP_GPT_MODEL_SUBHEAD_LOCATION=EXAMPLE KEY \
  -e APP_GPT_MODEL_SUBHEAD_HEADING_LOCATION=EXAMPLE KEY \
  -e APP_GPT_MODEL_TESTIMONIAL=EXAMPLE KEY \
  -e APP_GPT_MODEL_SERVICE_NAME=EXAMPLE KEY \
  -e APP_GPT_MODEL_SERVICE_DESCRIPTION=EXAMPLE KEY \
  -e APP_GPT_MODEL_TRANSLATION=EXAMPLE KEY \
  -e APP_GPT_MODEL_FROM_REVIEWS_TO_BUSINESS_DESC=EXAMPLE KEY \
  -e APP_GPT_MODEL_BUSINESS_NAMES=EXAMPLE KEY \
  -e APP_GPT_MODEL_BUSINESS_TYPE=EXAMPLE KEY \
  -e APP_GOOGLE_PLACES_API_KEY=EXAMPLE KEY \
  -e APP_FACEBOOK_ID=EXAMPLE KEY \
  -e APP_FACEBOOK_CONFIG_ID=EXAMPLE KEY \
  -e APP_FACEBOOK_SECRET=EXAMPLE KEY \
  -e APP_FACEBOOK_API_VERSION=v23.0 \
  -e APP_UNSPLASH_TRAFFIC_SOURCE_URL=http://example.com \
  -e APP_UNSPLASH_API_KEY=EXAMPLE KEY \
  -e APP_GET_TEMPLATE_PROJECT_URL=https://template-mk.b-cdn.net/api/ai-project-template \
  -e APP_GET_TEMPLATE_PAGE_URL=https://template-mk.b-cdn.net/api/ai-page-template \
  -e APP_GET_JSON_HTML_TEMPLATES_URL=https://template-mk.b-cdn.net/api/ai-list-template \
  -e APP_GET_JSON_HTML_STYLES_URL=https://template-mk.b-cdn.net/api/ai-list-styles \
  -e APP_VALIDATE_API_KEY_URL=http://example.com/validate \
  -e SENTRY_LARAVEL_DSN= \
  -e SENTRY_TRACES_SAMPLE_RATE= \
  -e APP_HTTP_SCHEME=http \
  brizy/ai:latest

Option 2: Docker Compose

Create a docker-compose.yml file with all environment variables:

version: '3.8'
services:
  brizy-ai:
    image: brizy/ai:latest
    container_name: brizy-ai
    ports:
      - "80:80"
    environment:
      # Base Configuration For developmnt
      APP_URL: http://localhost
      APP_ENV: dev
      APP_ENCODE_KEY: base64:dGVzdA==
      APP_HTTP_SCHEME: http
      
      # AWS S3 Configuration
      AMAZON_S3_REGION: Bucket Region
      AMAZON_S3_BUCKET: Bucket Name
      AMAZON_S3_KEY: EXAMPLE KEY
      AMAZON_S3_SECRET: EXAMPLE KEY

      # OpenAI Fine-tuned Models
      # Note: Brizy provides these pre-trained models. Contact support for model IDs.
      APP_GPT_MODEL_HEADING: EXAMPLE KEY
      APP_GPT_MODEL_HEADING_LOCATION: EXAMPLE KEY
      APP_GPT_MODEL_SUBHEAD: EXAMPLE KEY
      APP_GPT_MODEL_SUBHEAD_HEADING: EXAMPLE KEY
      APP_GPT_MODEL_SUBHEAD_LOCATION: EXAMPLE KEY
      APP_GPT_MODEL_SUBHEAD_HEADING_LOCATION: EXAMPLE KEY
      APP_GPT_MODEL_TESTIMONIAL: EXAMPLE KEY
      APP_GPT_MODEL_SERVICE_NAME: EXAMPLE KEY
      APP_GPT_MODEL_SERVICE_DESCRIPTION: EXAMPLE KEY
      APP_GPT_MODEL_TRANSLATION: EXAMPLE KEY
      APP_GPT_MODEL_FROM_REVIEWS_TO_BUSINESS_DESC: EXAMPLE KEY
      APP_GPT_MODEL_BUSINESS_NAMES: EXAMPLE KEY
      APP_GPT_MODEL_BUSINESS_TYPE: EXAMPLE KEY
      
      # Google Places API
      APP_GOOGLE_PLACES_API_KEY: KEY
      
      # Facebook API
      APP_FACEBOOK_API_VERSION: v23.0
      APP_FACEBOOK_ID: EXAMPLE KEY
      APP_FACEBOOK_CONFIG_ID: EXAMPLE KEY
      APP_FACEBOOK_SECRET: EXAMPLE KEY
      
      # Unsplash API
      APP_UNSPLASH_TRAFFIC_SOURCE_URL: http://exmple.com
      APP_UNSPLASH_API_KEY: EXAMPLE KEY
      
      # Template URLs
      APP_GET_TEMPLATE_PROJECT_URL: https://template-mk.b-cdn.net/api/ai-project-template
      APP_GET_TEMPLATE_PAGE_URL: https://template-mk.b-cdn.net/api/ai-page-template
      APP_GET_JSON_HTML_TEMPLATES_URL: https://template-mk.b-cdn.net/api/ai-list-template
      APP_GET_JSON_HTML_STYLES_URL: https://template-mk.b-cdn.net/api/ai-list-styles
      APP_VALIDATE_API_KEY_URL: http://example.com/validate
      
      # Sentry (Optional)
      SENTRY_LARAVEL_DSN: ""
      SENTRY_TRACES_SAMPLE_RATE: ""
    restart: unless-stopped

Then run:

docker-compose up -d

Fine-Tuned Models

Training Your Own Fine-Tuned Models

The OpenAI GPT model IDs (starting with APP_GPT_MODEL_*) reference fine-tuned models that you train using Brizy's provided prompt files for optimal website content generation. These models should be trained for:

Business headings and subheadings
Location-specific content
Testimonials and reviews
Service names and descriptions
Business name generation
Multi-language translation
Industry-specific content

Training Process:

Contact Brizy support to receive the training prompt files for each content type
Use OpenAI's fine-tuning API to train models with the provided prompt datasets
Update environment variables with your trained model IDs (e.g., ft:gpt-4o-mini-2024-07-18:your-org:model-name:id)

How to Fine-Tune Models

Once you receive the prompt files from Brizy, you'll need to fine-tune models using OpenAI's API. Here's a quick example:

Step 1: Prepare your training data

# Brizy provides JSONL files like:
# - heading_training.jsonl
# - testimonial_training.jsonl
# - service_description_training.jsonl
# etc.

Step 2: Upload training file to OpenAI

curl https://api.openai.com/v1/files \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -F "purpose=fine-tune" \
  -F "file=@heading_training.jsonl"

Step 3: Create fine-tuning job

curl https://api.openai.com/v1/fine_tuning/jobs \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "training_file": "file-abc123",
    "model": "gpt-4o-mini-2024-07-18"
  }'

Step 4: Get your fine-tuned model ID

# After training completes, you'll receive a model ID like:
# ft:gpt-4o-mini-2024-07-18:your-org:heading-model:abc123

Resources:

Note: You'll need to repeat this process for each content type (headings, testimonials, services, etc.) and update the corresponding environment variables with your trained model IDs.

Environment Variables Reference

Variable	Description	Example
APP_URL	The public base URL where Brizy AI will be accessible. Set this to the domain or IP where users will access the service.	`https://example.com`
APP_ENV	Environment mode (`dev` or `prod`). Use `dev` in development and `prod` in production for security.	`dev`
APP_ENCODE_KEY	Secret key used for encrypting sensitive data (e.g., cookies, tokens). Must start with `base64:` followed by your base64-encoded key. Should be a secure, random string. Never share or expose this value.	`base64:dGVzdA==`
APP_HTTP_SCHEME	Protocol scheme for the app (`http` or `https`). Use `https` in production for security.	`http`
AMAZON_S3_REGION	AWS region where your S3 bucket is hosted. Must match your actual AWS S3 region.	`region`
AMAZON_S3_BUCKET	Name of the AWS S3 bucket used for storing user assets, uploads, and configuration files.	`buket name`
AMAZON_S3_KEY	AWS access key ID for S3 access. Should have permissions for the specified bucket.	`1234`
AMAZON_S3_SECRET	AWS secret access key for S3 access. Keep this value secure and never commit to version control.	`1234`
APP_GPT_MODEL_HEADING	Fine-tuned OpenAI model for generating business headings. Change only if you have a custom model.	`ft:gpt-4o-mini-2024-07-18:1234`
APP_GPT_MODEL_HEADING_LOCATION	Fine-tuned model for generating location-specific headings. Useful for local businesses.	`ft:gpt-4o-mini-2024-07-18:1234`
APP_GPT_MODEL_SUBHEAD	Model for generating subheadings. Used for section titles or taglines.	`ft:gpt-4o-mini-2024-07-18:1234`
APP_GPT_MODEL_SUBHEAD_HEADING	Model for subheadings that relate to a main heading.	`ft:gpt-4o-mini-2024-07-18:1234`
APP_GPT_MODEL_SUBHEAD_LOCATION	Model for subheadings with a location context.	`ft:gpt-4o-mini-2024-07-18:1234`
APP_GPT_MODEL_SUBHEAD_HEADING_LOCATION	Model for subheadings that combine heading and location context.	`ft:gpt-4o-mini-2024-07-18:1234`
APP_GPT_MODEL_TESTIMONIAL	Model for generating customer testimonials. Useful for social proof sections.	`ft:gpt-4o-mini-2024-07-18:1234`
APP_GPT_MODEL_SERVICE_NAME	Model for generating names of services offered by a business.	`ft:gpt-4o-mini-2024-07-18:1234`
APP_GPT_MODEL_SERVICE_DESCRIPTION	Model for generating descriptions of business services.	`ft:gpt-4o-mini-2024-07-18:1234`
APP_GPT_MODEL_TRANSLATION	Model for translating content into different languages. Set to your custom translation model if needed.	`ft:gpt-4o-mini-2024-07-18:1234`
APP_GPT_MODEL_FROM_REVIEWS_TO_BUSINESS_DESC	Model for generating business descriptions from customer reviews.	`ft:gpt-4o-mini-2024-07-18:1234`
APP_GPT_MODEL_BUSINESS_NAMES	Model for generating business names.	`ft:gpt-4o-mini-2024-07-18:1234`
APP_GPT_MODEL_BUSINESS_TYPE	Model for classifying or suggesting business types.	`ft:gpt-4o-mini-2024-07-18:1234`
APP_GOOGLE_PLACES_API_KEY	API key for Google Places integration. Enables location search and autocomplete features. Obtain from Google Cloud Console.	`1234`
APP_FACEBOOK_ID	Facebook ID for Facebook integration.	`1234`
APP_FACEBOOK_CONFIG_ID	Facebook Config ID for Facebook integration.	`1234`
APP_FACEBOOK_SECRET	Facebook Secret for Facebook integration.	`1234`
APP_FACEBOOK_API_VERSION	Facebook api version for Facebook integration.	`v23.0`
APP_UNSPLASH_API_KEY	API key for Unsplash image search. Required for fetching stock images. Obtain from Unsplash Developers portal.	`1234`
APP_UNSPLASH_TRAFFIC_SOURCE_URL	The URL reported to Unsplash as the traffic source. Set to your app or company website.	`https://example.com`
APP_GET_TEMPLATE_PROJECT_URL	Endpoint to fetch project templates (JSON). Used for project scaffolding.	`https://template-mk.b-cdn.net/api/ai-project-template`
APP_GET_TEMPLATE_PAGE_URL	Endpoint to fetch page templates (JSON). Used for page generation.	`https://template-mk.b-cdn.net/api/ai-page-template`
APP_GET_JSON_HTML_TEMPLATES_URL	Endpoint to fetch all HTML preview templates (JSON). Used for template previews.	`https://template-mk.b-cdn.net/api/ai-list-template`
APP_GET_JSON_HTML_STYLES_URL	Endpoint to fetch all global style templates (JSON). Used for style/theme selection.	`https://template-mk.b-cdn.net/api/ai-list-styles`
APP_VALIDATE_API_KEY_URL	Endpoint to validate API keys. Used for internal or admin validation of API access.	`https://example.com/validate-key`
SENTRY_LARAVEL_DSN	DSN for Sentry error tracking. Leave empty if not using Sentry.	`""`
SENTRY_TRACES_SAMPLE_RATE	Sentry traces sample rate (0-1). Controls how much tracing data is sent.	`""`

Next Steps

After you get Brizy AI running with Docker, follow the API Reference to understand the complete flow and API endpoints for integration.

Quick Setup​

Option 1: Docker Run Command​

Option 2: Docker Compose​

Fine-Tuned Models​

How to Fine-Tune Models​

Environment Variables Reference​

Next Steps​