Configuration Constants

All settings can be configured via the Talking Head > Settings admin page. You can also override any setting using PHP constants in wp-config.php or environment variables of the same name.

Priority: constant > environment variable > database (admin UI) > default

When a setting is locked by a constant or environment variable, the admin field is disabled and shows “Locked by constant or environment variable.”

Provider

// Which TTS provider to use: 'openai', 'azure_openai', or 'wordpress' (WP 7.0+)
define( 'TALKING_HEAD_TTS_PROVIDER', 'openai' );

// OpenAI API key
define( 'TALKING_HEAD_OPENAI_API_KEY', 'sk-...' );

// TTS model: 'tts-1', 'tts-1-hd', or 'gpt-4o-mini-tts'
define( 'TALKING_HEAD_OPENAI_TTS_MODEL', 'tts-1' );

// Default voice for new speaker profiles: 'alloy', 'echo', 'fable', 'onyx', 'nova', or 'shimmer'
define( 'TALKING_HEAD_DEFAULT_VOICE', 'alloy' );

Azure OpenAI

// Azure OpenAI API key
define( 'TALKING_HEAD_AZURE_OPENAI_API_KEY', '...' );

// Azure OpenAI endpoint URL
define( 'TALKING_HEAD_AZURE_OPENAI_ENDPOINT', 'https://my-resource.openai.azure.com/' );

// Azure OpenAI deployment ID (name of your TTS deployment)
define( 'TALKING_HEAD_AZURE_OPENAI_DEPLOYMENT_ID', 'tts-hd' );

// Azure OpenAI API version
define( 'TALKING_HEAD_AZURE_OPENAI_API_VERSION', '2024-05-01-preview' );

Audio Processing

// Stitching mode: 'file' (concatenate on server) or 'virtual' (serve segments individually)
define( 'TALKING_HEAD_STITCHING_MODE', 'file' );

// Absolute path to the FFmpeg binary (optional — PHP fallback if not found)
define( 'TALKING_HEAD_FFMPEG_PATH', '/usr/bin/ffmpeg' );

// Output format: 'mp3' or 'aac'
define( 'TALKING_HEAD_OUTPUT_FORMAT', 'mp3' );

// Output bitrate: '128k', '192k', '256k', or '320k'
define( 'TALKING_HEAD_OUTPUT_BITRATE', '192k' );

// Milliseconds of silence between turns (0–5000)
define( 'TALKING_HEAD_SILENCE_GAP_MS', '500' );

Limits

// Maximum turns per episode (1–200)
define( 'TALKING_HEAD_MAX_SEGMENTS', '50' );

// Maximum characters per turn (100–4096)
define( 'TALKING_HEAD_MAX_SEGMENT_CHARS', '4096' );

// API requests per minute (1–60)
define( 'TALKING_HEAD_RATE_LIMIT', '10' );

Reference Table

Constant	Default	Allowed Values
TALKING_HEAD_TTS_PROVIDER	openai	openai, azure_openai, wordpress
TALKING_HEAD_OPENAI_API_KEY	(empty)	Any string
TALKING_HEAD_OPENAI_TTS_MODEL	tts-1	tts-1, tts-1-hd, gpt-4o-mini-tts
TALKING_HEAD_DEFAULT_VOICE	alloy	alloy, echo, fable, onyx, nova, shimmer
TALKING_HEAD_AZURE_OPENAI_API_KEY	(empty)	Any string
TALKING_HEAD_AZURE_OPENAI_ENDPOINT	(empty)	Valid URL
TALKING_HEAD_AZURE_OPENAI_DEPLOYMENT_ID	(empty)	Any string
TALKING_HEAD_AZURE_OPENAI_API_VERSION	2024-05-01-preview	Any string
TALKING_HEAD_STITCHING_MODE	file	file, virtual
TALKING_HEAD_FFMPEG_PATH	/opt/homebrew/bin/ffmpeg	Any executable path
TALKING_HEAD_OUTPUT_FORMAT	mp3	mp3, aac
TALKING_HEAD_OUTPUT_BITRATE	192k	128k, 192k, 256k, 320k
TALKING_HEAD_SILENCE_GAP_MS	500	0–5000
TALKING_HEAD_MAX_SEGMENTS	50	1–200
TALKING_HEAD_MAX_SEGMENT_CHARS	4096	100–4096
TALKING_HEAD_RATE_LIMIT	10	1–60

📦 Source: soderlind/talking-head · Edit on GitHub