tts-accelerator 1.0.3

Real-time speech generation from short to massive prompts in a single request — seamless and scalable, regardless of the length.

TTS Accelerator

Real-time natural speech generation — delivers real-time, natural speech from text in as little as 2-3 seconds, handling even extremely long sentences (16,000+ words) with instant voice output—no waiting or file saving.

This library works entirely in RAM to make speech super fast using the Edge-TTS engine. Edge-TTS sounds great, but it slows way down when you ask it to read long sentences—that was a big bottleneck for real-time apps. TTS Accelerator fixes that: it starts talking in just 2–3 seconds, even for thousands of words, by streaming audio straight from memory with no file saves or delays. Whether you’re building a chatbot or powering my AI assistant Nisha, who needs instant replies, this tool makes your app feel quick, smart, and alive.

  

Want to ask questions or share ideas?
👉   — we're active and friendly!

💡 Actively Developed & Maintained: This project is continuously improved with regular updates, bug fixes, and support. You're never alone — the journey keeps moving forward! 🚀🔧

Feel free to reach out, open issues, or suggest ideas — we’re building this together. 🤝

📧 Ranjit: [e-mail]

📧 Uday: [e-mail]

---

What Problem It Solves❓

Most TTS systems force you to wait while they generate the entire audio file — especially for long texts — before playback starts.

That’s slow, frustrating, and unusable for real-time apps.

TTS Accelerator solves this by streaming audio in real time.
It starts playback in just 2–3 seconds, even for huge texts. No disk I/O, no waiting — just seamless, instant voice output.
Perfect for:

• Voice assistants
• Chatbots
• Narration tools
• Live reading apps

---

📖 Quick Index

• 💡 Insight
• ✨ Key Features
• ❓ What Problem It Solves
• 🧠Why we Made This Library
• ⬇️ Installation
• 📚 Examples
• 📦 Requirements
• 📖 Documentation
• 👨‍🎓 Creators
• 📝 License (MIT)

💡Insight

TTS Accelerator is designed for ultra-fast playback. Unlike typical TTS systems that wait for full audio generation, this one begins speaking within 2–3 seconds — even for texts up to 16,000+ words. It streams while generating, offering smooth, natural output with zero delay.

✨Key Features

• Super Fast Speech (Real-Time TTS): Begins speaking in just 2–3 seconds, even for extremely long texts — up to 16,000+ words.
• Highly Compatible: Seamlessly works with popular TTS engines like Edge-TTS (tested), Coqui TTS etc..
• Streaming Playback: Audio starts playing instantly while it’s still being generated — no delay, no wait.
• Minimal API: Just a single function call to start speaking — speak_text(text).
• Zero-Delay Playback: Audio plays with zero delay, It is a smart system that uses fast RAM-based buffering and a separate producer-consumer pipeline to keep things smooth and super quick.

🧠Why we Made This Library

While working on our personal AI assistant, Nisha, we (Uday Subba & Ranjit Das) ran into a huge bottleneck:
Most TTS libraries either sounded robotic or took forever to generate long sentences — especially when trying to keep the voice natural.

We needed something fast enough for real-time conversation but still clear and human-like — and nothing we found was good enough.
That’s why we built TTS Accelerator.

It streams audio from text in just seconds, even for huge inputs, and it feels natural — exactly what Nisha needed to sound like a confident and responsive AI.

So yes — this library is not just an experiment, it powers a bigger dream.

---

Curious about my AI assistant Nisha?
She's the reason this library was born — and she's coming soon.
Stay tuned on GitHub or Discord for updates!

 

🔧Installation

You can install it from PyPI by running the following command:

pip install tts-accelerator

Or you can install it from source:

pip install .

or

git clone https://github.com/RanjitDas-IN/tts-accelerator.git
cd tts-accelerator/example_uses
python basic_uses.py

Explore the TTS Accelerator Source Code

📚Examples

Simple Usage:

# Import the library
import tts_accelerator as tts

# Example long text to demonstrate real-time speech generation.
text = (
    "Imagine reading out a 1,000-word story or a chatbot message stream — "
    "normally, you'd wait several seconds or even minutes before hearing anything. "
    "But with tts-accelerator, audio playback begins in just 2–3 seconds, "
    "no matter how long the input is. It streams audio directly from RAM, "
    "without saving to disk, and keeps the voice natural and fluid throughout. "
    "This makes it perfect for assistants, narrators, or any real-time voice-based apps."
)

# Speak the text — playback starts almost instantly
tts.speak_text(text)
# it will generate the audio in less than 3 seconds regardless of number of lines in the 'text variable'

📦Requirements

• Python v3.8+
• A stable internet connection

📖Documentation

https://tts-accelerator.readthedocs.io/

🤝 Support & Contribution

Have questions or ideas to improve the project? We’re building this together!

• Discord Support: Join our active community
• Open Issues: GitHub Issues Page
• Source Code: GitHub Repository
• Email: [e-mail]

We welcome bug reports, feature requests, or maybe a quick hello if you're enjoying the project!

Creators

Built and maintained by Uday Subba & Ranjit Das

Made with passion and precision.

📝License

This project is licensed under MIT.

CTA:

Try it now and tell us what you build!

🔮 About Nisha (Coming Soon)

Nisha is my personal AI assistant — sassy, smart, and ultra-fast.
She’s designed to handle real-time conversations with confidence and clarity, and tts-accelerator is one of the core techs that power her voice.

Want to follow her journey or be the first to try her out?

• GitHub Updates: RanjitDas-IN
• Chat on Discord: Join the community