Credits & Acknowledgements#

Hey! This section is all about giving credit where it’s due - both to me for building this, and to all the amazing open-source libraries and tools that made it possible.

Who Built This#

Aman Kumar Pandey (that’s me!)

I’m the person behind Vaani. I designed it, wrote the code, debugged it at 2 AM, and kept working on it because I genuinely found it interesting.

Some background about me: - I love working on AI and voice technology - Built this to solve my own frustration with existing assistants - Big believer in privacy-first technology - Still learning and improving this project

Find me at: - GitHub: @paman7647 - Email: paman7647@gmail.com

Why I Made This#

I wanted to build something that:

  1. Respects Privacy: I don’t like the idea of all my voice commands going to some company’s servers

  2. Actually Works Offline: Living in India, internet isn’t always reliable - needed something that works regardless

  3. Supports Multiple Languages: I speak Hindi and English daily, wanted my assistant to do the same

  4. Is Open and Customizable: I like understanding and modifying my tools

  5. Teaches Me New Things: This project has been an incredible learning experience

Plus, it’s just really cool to have your own AI assistant that you built yourself!

Built With These Awesome Tools#

I couldn’t have done this alone. Here are all the open-source projects that power Vaani:

The Brain (AI & Recognition)

  • Google Gemini: This powers the conversational AI. Makes Vaani actually smart and able to have real conversations. Thanks Google for the API!

  • Vosk: My favorite - offline speech recognition that’s fast and accurate. Works when internet is down.

  • PocketSphinx: The backup system. Not the best, but always there when needed.

Audio Handling

  • PyAudio: Handles all the microphone input and speaker output. Cross-platform magic.

  • SpeechRecognition: Makes it easy to work with multiple speech engines without going crazy.

Playing Music & Speaking

  • VLC (libvlc): Plays music from YouTube reliably. VLC is just awesome for media playback.

  • PyTTSx3: Converts text to speech offline. Voices aren’t perfect but they work!

  • yt-dlp: Extracts audio from YouTube. This tool is incredible.

Helper Libraries

  • RapidFuzz: Super fast fuzzy string matching. Perfect for wake word detection.

  • DuckDuckGo Search: Privacy-friendly web search. No tracking, just results.

  • BeautifulSoup4: When I need to parse web pages and extract information.

  • python-dotenv: Manages environment variables and API keys securely.

  • colorama: Makes terminal output look nice with colors.

Want to Contribute?#

I’d love your help! If you want to contribute:

  • Found a bug? Open an issue on GitHub

  • Have an idea? Start a discussion

  • Want to add code? Check out the Contributing Guide

  • Just using it? Star the repo - it motivates me!

GitHub: paman7647/vaani

Thank You#

Seriously, thanks to everyone who:

  • Uses Vaani and gives feedback

  • Reports bugs (helps me make it better)

  • Contributes code or documentation

  • Stars the repo (it’s encouraging!)

  • Shares ideas for improvements

And huge thanks to the open-source community. Without these libraries and tools, projects like Vaani wouldn’t be possible for individual developers like me.

Special shoutout to anyone who’s ever answered my Stack Overflow questions or written helpful tutorials. You’re the real MVPs!