Text to Speech Guide

pip install fish-audio-sdk

from fish_audio_sdk import Session, TTSRequest

# Initialize session
session = Session("your_api_key")

# Generate speech
with open("output.mp3", "wb") as f:
    for chunk in session.tts(TTSRequest(
        text="Hello, world!",
        reference_id="your_voice_model_id"
    )):
        f.write(chunk)

# Use a voice from the playground
session.tts(TTSRequest(
    text="Welcome to Fish Audio!",
    reference_id="7f92f8afb8ec43bf81429cc1c9199cb1"
))

# Use your own cloned voice
session.tts(TTSRequest(
    text="This is my custom voice speaking",
    reference_id="your_model_id"
))

from fish_audio_sdk import ReferenceAudio

# Use reference audio on-the-fly
session.tts(TTSRequest(
    text="Hello from reference audio",
    references=[
        ReferenceAudio(
            audio=open("voice_sample.wav", "rb").read(),
            text="Sample text from the audio"
        )
    ]
))

# Using the latest S1 model
session.tts(
    TTSRequest(text="Hello world"),
    backend="s1"
)

session.tts(TTSRequest(
    text="Your text here",
    format="mp3",  # Options: "mp3", "wav", "pcm"
    mp3_bitrate=128  # For MP3: 64, 128, or 192
))

session.tts(TTSRequest(
    text="Long text content...",
    chunk_length=200  # 100-300 characters per chunk
))

session.tts(TTSRequest(
    text="Quick response needed",
    latency="balanced"  # "normal" or "balanced"
))

import httpx
import ormsgpack

# Prepare request
request_data = {
    "text": "Hello, world!",
    "reference_id": "your_model_id",
    "format": "mp3"
}

# Make API call
with httpx.Client() as client:
    response = client.post(
        "https://api.fish.audio/v1/tts",
        content=ormsgpack.packb(request_data),
        headers={
            "authorization": "Bearer YOUR_API_KEY",
            "content-type": "application/msgpack",
            "model": "s1"
        }
    )
    
    # Save audio
    with open("output.mp3", "wb") as f:
        f.write(response.content)

# Stream audio chunks
with open("stream_output.mp3", "wb") as f:
    for chunk in session.tts(TTSRequest(
        text="Streaming this text in real-time",
        reference_id="model_id"
    )):
        f.write(chunk)
        # Process chunk immediately for real-time playback

# Add emotion markers to your text
emotional_text = """
(excited) I just won the lottery!
(sad) But then I lost the ticket.
(laughing) Just kidding, I found it!
"""

session.tts(TTSRequest(
    text=emotional_text,
    reference_id="model_id"
))

texts = [
    "First announcement",
    "Second announcement",
    "Third announcement"
]

for i, text in enumerate(texts):
    with open(f"output_{i}.mp3", "wb") as f:
        for chunk in session.tts(TTSRequest(
            text=text,
            reference_id="model_id"
        )):
            f.write(chunk)

import time

def generate_with_retry(text, max_retries=3):
    for attempt in range(max_retries):
        try:
            audio_data = b""
            for chunk in session.tts(TTSRequest(
                text=text,
                reference_id="model_id"
            )):
                audio_data += chunk
            return audio_data
        except Exception as e:
            if attempt < max_retries - 1:
                time.sleep(2 ** attempt)  # Exponential backoff
            else:
                raise e