TTS(1)

tts - text-to-speech converter using ElevenLabs API by jtgi

INSTALLATION

curl -fsSL https://raw.githubusercontent.com/jtgi/tts/main/install.sh | sh

EXAMPLES

Convert simple text to speech

echo "Hello world" | tts > hello.mp3

Process a large book with custom chunk size

cat book.txt | tts -c 10000 > book.mp3

Play audio directly (macOS only)

echo "Hello" | tts --play

Extract and convert chapter headings

grep "^Chapter" book.txt | tts -v EXAVITQu4vr4xnSDxMaL > chapters.mp3

SYNOPSIS

tts [OPTIONS] < input > output.mp3

OPTIONS:
  -k, --api-key KEY         API key
  -v, --voice ID            Voice ID
  -m, --model ID            Model ID  
  -f, --format FORMAT       Output format
  -c, --chunk-size SIZE     Chunk size
  -d, --debug              Debug mode
  -p, --play               Play directly
  -l, --latency LEVEL      Latency optimization
  --stability FLOAT        Voice stability
  --similarity-boost FLOAT Similarity boost
  --style FLOAT            Style intensity
  --speaker-boost BOOL     Speaker boost
  --seed INTEGER           Reproducible seed
  --previous-text TEXT     Previous context
  --next-text TEXT         Next context
  --previous-request-ids   Previous IDs
  --next-request-ids       Next IDs

DESCRIPTION

tts reads text from standard input and writes audio to standard output. It uses the ElevenLabs API to convert text to speech, processing input in chunks for streaming large files.

OPTIONS

-k, --api-key KEY

ElevenLabs API key. If not specified, uses the ELEVENLABS_API_KEY environment variable.

-v, --voice ID

Voice ID to use for speech synthesis. Default: 21m00Tcm4TlvDq8ikWAM (Rachel)

-m, --model ID

Model ID to use for speech synthesis. Default: eleven_monolingual_v1

-f, --format FORMAT

Output audio format. Default: mp3_44100_128

-c, --chunk-size SIZE

Number of characters to process per chunk. Useful for large files. Default: 5000

-d, --debug

Enable debug output to stderr. Shows API requests, responses, and detailed error information.

-p, --play

Play audio directly without saving (macOS only).

-l, --latency LEVEL

Optimize streaming latency (0-4, higher = lower latency). Default: 1

--stability FLOAT

Voice stability (0.0-1.0). Higher values make voice more consistent but less expressive. Default: 0.5

--similarity-boost FLOAT

Voice similarity boost (0.0-1.0). Enhances similarity to the original voice. Default: 0.75

--style FLOAT

Speaking style intensity (0.0-1.0). Controls how expressive the voice is. Default: 0.0

--speaker-boost BOOL

Enable speaker boost (true/false). Enhances voice clarity and distinction. Default: true

--seed INTEGER

Seed for reproducible generation. Use the same seed to get identical audio for the same text.

--previous-text TEXT

Previous text context for better pronunciation and intonation.

--next-text TEXT

Next text context for better pronunciation and intonation.

--previous-request-ids IDS

Comma-separated previous request IDs for context continuity.

--next-request-ids IDS

Comma-separated next request IDs for context continuity.

-h, --help

Display help message and exit.

ENVIRONMENT

`ELEVENLABS_API_KEY`	Default API key for ElevenLabs service
`TTS_VOICE_ID`	Default voice ID
`TTS_MODEL_ID`	Default model ID
`TTS_OUTPUT_FORMAT`	Default output format
`TTS_DEBUG`	Set to 1 to enable debug mode by default
`TTS_CHUNK_SIZE`	Default chunk size for processing
`TTS_OPTIMIZE_STREAMING_LATENCY`	Default latency optimization level (0-4)
`TTS_STABILITY`	Default voice stability (0.0-1.0)
`TTS_SIMILARITY_BOOST`	Default similarity boost (0.0-1.0)
`TTS_STYLE`	Default speaking style intensity (0.0-1.0)
`TTS_USE_SPEAKER_BOOST`	Default speaker boost setting (true/false)
`TTS_SEED`	Default seed for reproducible generation
`TTS_PREVIOUS_TEXT`	Default previous text context
`TTS_NEXT_TEXT`	Default next text context
`TTS_PREVIOUS_REQUEST_IDS`	Default previous request IDs
`TTS_NEXT_REQUEST_IDS`	Default next request IDs

VOICES

Common voice IDs include:

`21m00Tcm4TlvDq8ikWAM`	Rachel (default)
`EXAVITQu4vr4xnSDxMaL`	Bella
`ErXwobaYiN019PkySvjV`	Antoni
`MF3mGyEYCl7XYWbV9V6O`	Elli
`TxGEqnHWrfWFTfGW9XjX`	Josh

FORMATS

`mp3_44100_128`	MP3 128kbps (default)
`mp3_44100_192`	MP3 192kbps
`pcm_16000`	PCM 16kHz
`pcm_22050`	PCM 22kHz
`pcm_24000`	PCM 24kHz
`pcm_44100`	PCM 44.1kHz
`ulaw_8000`	μ-law 8kHz

EXIT STATUS

Success

Error occurred (API failure, missing input, authentication error, etc.)

NOTES

Requires curl(1) to be installed.

The tool processes text in chunks to handle large files efficiently, making it suitable for converting entire books or long documents.

API errors are reported with HTTP status codes and helpful hints for common issues like authentication failures or rate limiting.