This project aims to develop a local voice-to-text pipeline stack that converts spoken language into transcribed text, feeding it into an AI chatbot interface displayed on a browser. The system operates entirely locally and without relying on external servers or services, utilizing open source tools for efficient resource utilization. ๐ป
- Speech-to-Text conversion: Convert spoken language into transcribed text using OpenAI's Whisper API. ๐ค
- Embeddings storage and chat history management: Utilize embeddings to store reference documents and manage separate embeddings for the current chat history. ๐
- Local pipeline stack: Operate entirely locally, without relying on external servers or services. ๐
- Offline security feature: Leverage the system's offline nature as a security measure. ๐ก๏ธ
- Privacy protection: Automatically delete logs at the end of each session to maintain user privacy. ๐ซ
- Custom implementation: With the extension running in Developer Mode and not intended to be distributed on the Chrome Web Store, this project will not be restrained by Chrome Extension policies. ๐ก๏ธ
This project is in its infancy, but we have planned out a blueprint that attempts to capture all of the above features. More details will be added as the development progresses. ๐
- Aidain: AI Senior Project Developer specializing in Computer Science, Programming, and open source AI initiatives. ๐ป
- Les-El: Human amateur programmer and developer, guiding Aidain and other AI tools. ๐จโ๐ป
We welcome all feedback, suggestions for improvements, and contributions to this project. Please feel free to create an issue or submit a pull request on our GitHub repo. Your help is greatly appreciated. ๐