Tools · SPBoucher
Menu

Whisper

Maintenance

by OpenAI · Model · also known as openai-whisper, OpenAI Whisper

OpenAI's open-source multilingual speech recognition model family with MIT-licensed code and weights, a Python package and CLI for transcription, translation and language identification

MIT licenseOpen sourceFreeLocalSelf-hostable
Visit official site Compare

Link verified Nov 2, 2026

Overview

Whisper is a general-purpose speech recognition model released by OpenAI in 2022 with code and weights under the MIT license. A single Transformer sequence-to-sequence model trained on a large, diverse audio corpus performs multilingual transcription, translation into English and language identification. Six sizes are published, from tiny to large, plus turbo, an optimized variant of large-v3 that trades a little accuracy for speed and is not trained for translation; English-only variants exist for the smaller sizes. The openai-whisper Python package exposes load_model and transcribe, and a whisper CLI handles files directly; dependencies are PyTorch, tiktoken and ffmpeg.

Whisper is the foundation of a large downstream ecosystem: faster-whisper and whisper.cpp reimplement inference for CPUs and small devices, mlx-whisper ports it to Apple Silicon, and OpenAI's hosted transcription API offers the same capability without local setup. The original repository changes infrequently and serves mainly as a reference, so projects needing speed or long-form robustness usually pick one of the derivatives. Accuracy varies substantially by language and audio quality, timestamps are approximate, and speaker diarization is not included.

Editorial note on uncertainty. product_status is set to maintenance because the repository receives infrequent updates and the README carries no roadmap; OpenAI has not declared it deprecated.

Capabilities

  • Multilingual transcription and translation to English
  • Automatic language identification
  • Six model sizes plus turbo for faster inference
  • Python API and CLI
  • MIT-licensed weights usable in derivative runtimes

Typical workflows

  1. Transcribe lecture recordings offline with the CLI
  2. Pick a model size to balance accuracy and speed on available hardware
  3. Swap to mlx-whisper or faster-whisper for production throughput

Strengths

  • Permissive MIT license for code and weights
  • Strong multilingual coverage
  • Large ecosystem of optimized reimplementations

Limitations

  • Reference implementation is slow compared with derivatives
  • No speaker diarization and approximate timestamps
  • Accuracy drops on low-resource languages and noisy audio

Prerequisites

  • Python with PyTorch and ffmpeg
  • GPU recommended for larger models

Who it is for

  • Researchers and students transcribing interviews or lectures
  • Developers embedding speech recognition offline

Alternatives & comparison

  • ElevenLabs

    Hosted voice AI platform: text-to-speech and speech-to-text models, voice cloning, dubbing, music and sound effects, a voice-agent builder and a REST API with Python and TypeScript SDKs

    ProprietaryFreemium
  • MLX Whisper

    Port of OpenAI's Whisper speech-recognition models to MLX, offering a CLI and Python API for local transcription on Apple Silicon

    Open sourceOpen source
  • OpenAI API

    OpenAI's hosted developer platform: Responses and Chat Completions APIs, Realtime voice, embeddings, images, audio, fine-tuning, evals and managed agents

    ProprietaryUsage-based
Compare Whisper with its alternatives →

Sources & verification

Primary sources consulted for this record. Claims are tied to the fields each source supports; the automated link checker re-verifies every URL on a schedule.

  • License
    https://github.com/openai/whisper/blob/main/LICENSE

    Supports: license_spdx, source_model

    checked Nov 1, 2026
  • Repository
    openai/whisper

    https://github.com/openai/whisper

    Supports: name, summary, capabilities, prerequisites

    The announcement page on openai.com returns 403 to automated clients; the repository is the canonical reference

    checked Nov 1, 2026
Confidence
High
Reviewed
Nov 1, 2026
Links
Verified · Nov 2, 2026

Spotted an error?

Broken link, outdated license, wrong category, or a tool that no longer exists? Tell us; reports are reviewed by the editor.

Gemini API

Google (Chrome DevTools team) · Model providers & APIs

Google's developer API for the Gemini model family, keyed from Google AI Studio, with function calling, Search grounding, Live API, image and video generation and a free tier

ProprietaryFreemiumUsage-basedNov 2, 2026

ComfyUI

Comfy Org · Voice, vision & multimodal AI

Open-source node-based interface and inference engine for diffusion and other generative models (image, video, audio, 3D), runnable locally with desktop apps or as a hosted cloud

GPL-3.0-onlyOpen sourceFreeLocalNov 2, 2026