API or local Whisper?

Use the API for convenience, local for privacy and cost control — the guide covers both.

How many languages does Whisper support?

Whisper supports roughly 100 languages for transcription and about 50 for translation to English. Accuracy varies by language — high-resource languages like English, Spanish, and Mandarin perform best; low-resource ones may need post-editing.

Which Whisper model size should I run?

For clean English, the small or base models are fast and accurate enough. Step up to medium or large for heavy accents, noisy audio, or non-English languages. On CPU the large model is slow — use a GPU, a faster reimplementation, or the hosted API when speed matters.

GuideIntermediate

OpenAI Whisper Speech Recognition Guide

Item: OpenAI Whisper Speech Recognition Guide
Rating: 4.5
Author: AI Resource Hub

Official documentation for OpenAI Whisper, the open-source speech recognition model trained on diverse multilingual audio. It covers installation, model size selection, and transcription API usage, with support for roughly 99 languages and direct translation to English. Aimed at developers who need accurate, locally runnable speech-to-text for subtitling, meeting notes, or voice search indexing.

Overview

"OpenAI Whisper Speech Recognition Guide" is a "Guide" resource curated by AI Resource Hub, filed under the Tutorials category and suited to Intermediate-level learners. It is provided by OpenAI, was last updated on 2026-06-16, and holds an editorial score of 4.5/5 from our team. Click "Visit Resource" on the right to open the original page.