OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Audio & Voice AI

Thewhisper

Thewhisper is a Python repository of optimized Whisper speech-recognition inference code and linked model weights, designed for streaming and on-device transcription on Apple Silicon and NVIDIA GPUs.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Thewhisper — OpenViST project profile in Audio & Voice AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when you need developer-facing Whisper inference code for streaming or on-device transcription and can provide a supported system and setup.

Who it is for

  • Developers adding live or file-based speech transcription to an application.
  • Teams building local macOS note-taking, captioning or voice-interface software.
  • Operators evaluating self-hosted inference on supported NVIDIA GPU hardware.

Documented uses

  • Transcribe audio files and return text, with optional word timestamps in the documented Apple example.
  • Process microphone audio as a stream for live transcription.
  • Build a local transcription application using the repository’s examples and documented local REST API.
  • Select a supported Apple Silicon or NVIDIA inference path for a transcription project.
Before you choose

Requirements & limits

The repository is public and identifies its license as MIT. The separate TheStage optimized-engine installation instructions require generating an access token on TheStage AI Platform.

  • The project is a repository and inference implementation, not a finished transcription service.
  • Apple Silicon and NVIDIA use different package extras, hardware requirements and inference paths.
  • The README lists streaming, multilingual operation, word timestamps and 10-, 15-, 20- and 30-second chunk modes for large-v3 and large-v3-turbo.
  • A separate optimized NVIDIA-engine route uses TheStage packages and requires a platform access token.
  • The repository links to open model weights for the large-v3-turbo model.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install Thewhisper for Apple Silicon

Set up the repository’s Apple-specific Python package path.

Before you start

  • Apple Silicon system listed by the project, with at least 2 GB RAM (4 GB recommended for large-v3).
  • macOS 15.0 or later as stated in the README.
  • Python 3.10–3.12.

Steps

  1. Clone the repository with git clone https://github.com/TheStageAI/TheWhisper.git.

  2. Change into the TheWhisper directory.

  3. Install the Apple package extras with pip install .[apple].

Expected outcome

The Apple-specific package installation step is completed; the README’s inference examples are the next stage for using a model.

Guide limitations

  • This recipe covers installation only, not running inference or building an application.
  • The README’s operating-system wording pairs macOS 15.0 with “Ventura”; confirm the system requirement for your target environment.
Documentation 1 ↗
Source-based practical guide

Install the standard NVIDIA package path

Install the repository’s NVIDIA Python package extras for a supported GPU environment.

Before you start

  • A supported NVIDIA GPU: RTX 4090, RTX 5090, L40s, H100, A100 or Jetson-Thor.
  • Ubuntu 20.04 or later, CUDA 11.8 or later, driver 520.0 or later, Python 3.10–3.12 and at least 2.5 GB RAM; 5 GB is recommended for large-v3.

Steps

  1. Clone the repository with git clone https://github.com/TheStageAI/TheWhisper.git.

  2. Change into the TheWhisper directory.

  3. Install the NVIDIA package extras with pip install .[nvidia].

Expected outcome

The standard NVIDIA package installation step is completed; the repository’s NVIDIA usage example documents the next inference stage.

Guide limitations

  • This recipe covers the standard NVIDIA installation, not the separate TheStage optimized-engine setup or execution of inference.
  • The listed GPU and software requirements are repository documentation, not a compatibility guarantee.
Documentation 1 ↗
COMPARE OPTIONS

Compare options for Model inference and serving

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
ThewhisperTranscribe audio files and return text, with optional word timestamps in the documented Apple example.macOS, Self-hostedThe repository is public and identifies its license as MIT. The separate TheStage optimized-engine installation instructions require generating an access token on TheStage AI Platform.
EdgedictDecode speech from a YouTube Live video into text.Cloud
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
13 October 2026
Original website title
TheStageAI/TheWhisper · Official homepage title · retained in its source language
How we review sourcesSuggest a correction