README.md

ReaSpeech

Painless speech-to-transcript inside of REAPER

ReaSpeech is a REAScript frontend that will take your project's media items and run them through a docker-hosted whisper backend, building a searchable, project-marker based index of the resulting transcription. That's right - not only do you reap (zing!) the benefits of a staff of transcribers, but you get to add "familiar with Docker" to your resume.

Quick Usage

• Install Docker
• Run the Docker image
• Navigate to localhost:9000

Docker Commands

CPU

docker run -d -p 9000:9000 techaudiodoc/reaspeech:latest

GPU (currently Windows/NVIDIA only)

docker run -d --gpus all -p 9000:9000 techaudiodoc/reaspeech:latest-gpu

For more detailed instructions, see Docker Usage

Documentation

• Docker Usage
• Running Outside of Docker
• Contributing
• Development

Credits

ReaSpeech Team

• Dave Benjamin
• Jason Nanna
• Kyle Vande Slunt
• Michael DeFreitas
• Roel Sanchez

Third-Party Software

ReaSpeech's initial web service and Docker setup were based on the Whisper ASR Webservice project.

Transcription is provided by the Faster Whisper library.

ReaSpeech uses ReaImGui by Christian Fillion for its user interface toolkit.

Licensing

ReaSpeech is licensed under the terms of the GPLv3. Portions are derived from the whisper-asr-webservice project, which is MIT-licensed. All source files in this repository should be considered GPL-licensed unless otherwise specified.