openai-tts

1# OpenAI Text-to-Speech
2 
3Generate high-quality spoken audio from text using OpenAI's TTS API.
4 
5## Authentication
6 
7The API key is available as environment variable:
8```bash
9OPENAI_API_KEY
10```
11 
12## Models
13 
14- `gpt-4o-mini-tts` - Newest, most reliable. Supports tone/style instructions.
15- `tts-1` - Lower latency, lower quality
16- `tts-1-hd` - Higher quality, higher latency
17 
18## Voice Options
19 
20Built-in voices (English optimized):
21- `alloy`, `ash`, `ballad`, `coral`, `echo`, `fable`
22- `nova`, `onyx`, `sage`, `shimmer`, `verse`
23- `marin`, `cedar` - **Recommended for best quality**
24 
25Note: `tts-1` and `tts-1-hd` only support: alloy, ash, coral, echo, fable, onyx, nova, sage, shimmer.
26 
27## Python Example
28 
29```python
30from pathlib import Path
31from openai import OpenAI
32 
33client = OpenAI()  # Uses OPENAI_API_KEY env var
34 
35# Basic usage
36with client.audio.speech.with_streaming_response.create(
37    model="gpt-4o-mini-tts",
38    voice="coral",
39    input="Hello, world!",
40) as response:
41    response.stream_to_file("output.mp3")
42 
43# With tone instructions (gpt-4o-mini-tts only)
44with client.audio.speech.with_streaming_response.create(
45    model="gpt-4o-mini-tts",
46    voice="coral",
47    input="Today is a wonderful day!",
48    instructions="Speak in a cheerful and positive tone.",
49) as response:
50    response.stream_to_file("output.mp3")
51```
52 
53## Handling Long Text
54 
55For long documents, split into chunks and concatenate:
56 
57```python
58from openai import OpenAI
59from pydub import AudioSegment
60import tempfile
61…