Configuration

Configure Subtide for your needs.

Extension Configuration

Click the Subtide extension icon to open the configuration popup.

Basic Settings

Setting	Description
Operation Mode	Select Tier 1-4 based on your needs
Backend URL	URL of your backend server (default: `http://localhost:5001`)
Target Language	Language to translate subtitles into

API Settings (Tier 1 & 2)

Setting	Description
API Provider	OpenAI, OpenRouter, or Custom Endpoint
API Key	Your API key for the selected provider
Model	LLM model to use for translation
API URL	Custom API endpoint (for custom providers)

Display Settings

Setting	Description
Subtitle Size	Small, Medium, Large, or XL
Dual Mode	Show both original and translated text

Backend Environment Variables

Configure the backend using environment variables or a .env file.

Core Settings

Variable	Description	Default
`PORT`	Server port	`5001`
`GUNICORN_WORKERS`	Number of worker processes	`2`
`GUNICORN_TIMEOUT`	Request timeout in seconds	`300`
`CORS_ORIGINS`	Allowed CORS origins	`*`

API Configuration (Tier 3/4)

Variable	Description	Default
`SERVER_API_KEY`	API key for LLM provider	—
`SERVER_API_URL`	LLM API endpoint	—
`SERVER_MODEL`	LLM model name	—

Whisper Configuration

Variable	Description	Default
`WHISPER_MODEL`	Model size: `tiny`, `base`, `small`, `medium`, `large-v3`, `large-v3-turbo`	`base`
`WHISPER_BACKEND`	Backend: `mlx`, `faster`, `openai`	Auto-detected

Example Configurations

Tier 1: Basic YouTube Translation

Extension:

Operation Mode: Tier 1 (Standard)
Backend URL: http://localhost:5001
API Provider: OpenAI
API Key: sk-...
Model: gpt-4o-mini
Target Language: Spanish

Backend:

./subtide-backend-macos

Tier 2: Whisper Transcription

Extension:

Operation Mode: Tier 2 (Enhanced)
Backend URL: http://localhost:5001
API Provider: OpenAI
API Key: sk-...
Model: gpt-4o
Target Language: Japanese

Backend:

WHISPER_MODEL=large-v3-turbo WHISPER_BACKEND=mlx ./subtide-backend-macos

Tier 3: Shared Server

Extension:

Operation Mode: Tier 3 (Managed)
Backend URL: https://your-server.com
Target Language: German

Backend:

SERVER_API_KEY=sk-xxx \
SERVER_API_URL=https://api.openai.com/v1 \
SERVER_MODEL=gpt-4o \
docker-compose up subtide-tier3

Tier 4: Streaming Mode

Extension:

Operation Mode: Tier 4 (Stream)
Backend URL: https://your-server.com
Target Language: French

Backend:

SERVER_API_KEY=sk-xxx \
SERVER_API_URL=https://api.openai.com/v1 \
SERVER_MODEL=gpt-4o \
WHISPER_MODEL=large-v3-turbo \
docker-compose up subtide-tier4

Local LLM with LM Studio

Extension:

Operation Mode: Tier 2 (Enhanced)
Backend URL: http://localhost:5001
API Provider: Custom Endpoint
API URL: http://localhost:1234/v1
API Key: lm-studio
Model: local-model
Target Language: Korean

Whisper Model Selection

Choose a model based on your hardware and quality needs:

Model	VRAM/RAM	Speed	Quality	Best For
`tiny`	~1 GB	Fastest	Basic	Testing
`base`	~1 GB	Fast	Good	General use
`small`	~2 GB	Medium	Better	Better accuracy
`medium`	~5 GB	Slow	Great	High quality
`large-v3`	~10 GB	Slowest	Best	Maximum quality
`large-v3-turbo`	~6 GB	Fast	Excellent	Recommended

Tip: Recommendation Use large-v3-turbo for the best balance of speed and quality.

Next Steps

YouTube Guide - Using Subtide with YouTube
Backend Overview - Backend deployment options