Whisper Transcriber

Overview

Whisper Transcriber transcribes audio into text using the amazing Whisper model from OpenAI. It supports both official OpenAI API and locally deployed Whisper models.

This simple tool is published for free for those who need to easily transcribe complex audio with OpenAI's Whisper models, all under your full control.

Features

• Voice recording, audio saving and loading
  
• Official OpenAI Transcription API
• Locally deployed Whisper model

Usage

You can now use Whisper Transcriber tool either with official OpenAI API, or with locally deployed Whisper models.

With OpenAI API

1. Register and obtain an OpenAI developer key from OpenAI (See video guide)
2. Open Transcriber
3. Type your key (If you wish this to be filled automatically, setup an environment variable OPENAI_API_KEY)
4. Select target language: see Dev Log for a full list of supported languages.
5. Start recording or load existing audio files to transcribe!

With Locally Deployed Models

1. Pick "Local" from dropdown to use locally deployed models.
2. Pick a suitable model. Larger models run slower but are more accurate. Smaller models generally suffice.
3. Pick a language or use "Auto" for automatic language detection.
4. When picking from available models, it will automatically download the models to the executable folder.
   
5. If you want to speed up the process, you can download those models from the download section of this webpage and place those model files yourself. See video guide for detail.

This software is cross-platform - Linux and Mac version is available on request. We are aware of the rather large size (170Mb) of this program and will try to reduce it in future releases.

Changelog

• v0.1: First build.
• v0.2: Component update per framework change; Make OpenAI key configurable during runtime.
• v0.3: Allow picking save path.
• v0.4: Improve GUI; Show recording durations; Update executable name.
• v.04.1: Support automatic segmentation; Improve HTTP request stability.
• v0.4.2: Add more language options. Mac build available for this version!
• v0.5: Support locally hosted model; Fix layout of status label. Adjust buttons layout. Expose OpenAI endpoint type.
• v0.5.1: Improve local model transcription performance; Reduce distributable folder size.

Disclaimer

THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

References

• Audio Transcription Guide: OpenAI Guide
• Official API: API Reference
• Whisper Model: Github

If you enjoyed this AI tool, please also consider our Chat Gateway service - which enables you to quickly add an AI-enabled chat bot to your personal website! Visit official website for more!