Open Source, Free Tier Capable Whispr Using Cloudflare AI (github.com)

Whispr, an innovative open-source voice-to-text tool, has been released, leveraging Cloudflare's AI capabilities to transcribe speech using Whisper and format the output with a Large Language Model (LLM). Users can simply press a hotkey, speak, and upon releasing the key, their spoken words are transcribed and automatically pasted into any text field, significantly enhancing productivity. The application streamlines the process by capturing audio via a Wails desktop app and processing it in the cloud through a Cloudflare Worker, enabling efficient speech-to-text transformation. This development is noteworthy for the AI/ML community as it exemplifies the integration of various AI technologies—cloud infrastructure, real-time audio processing, and LLMs—into a user-friendly application. The system operates at a sample rate of 16kHz, utilizing Whisper for speech-to-text transcription and an LLM for text formatting. With capabilities for customization and easy configuration, Whispr promises to simplify workflows for content creators, developers, and any professionals who require rapid note-taking and transcription, showcasing the potential of open-source solutions in enhancing daily tasks with AI.