VoxLingua107

by Tallinn University of Technology

6.6kh language-identification corpus — 107 languages from YouTube.

TL;DR

6.6kh language-identification corpus — 107 languages from YouTube.

Best for spoken language identification (LID) training and evaluation across 107 languages. Pricing: free.

Category
Open source
License
—
Stars
—
Last push
—
Pricing
free
Platforms
Web, HuggingFace

What it is

VoxLingua107 is 6.6k hours across 107 languages crawled from YouTube — the standard LID benchmark. License: CC BY 4.0 metadata.

Best for: Spoken language identification (LID) training and evaluation across 107 languages.
Watch out for: CC BY 4.0 metadata · YouTube TOS for audio · imbalanced per-language. Cite: Valk & Alumae, SLT 2021.

Install / use

http://bark.phon.ioc.ee/voxlingua107/

Features

Speaker diarizationNo
Word-level timestampsNo
Streaming / real-timeNo
Languages supported107
HIPAA eligibleNo

VoxLingua107 vs Whipscribe

FeatureVoxLingua107Whipscribe
CategoryOpen sourceTranscription APIs
Pricingfreefree beta
Speaker diarizationNoYes
Word timestampsNoYes
StreamingNoNo
Languages10799
PlatformsWeb, HuggingFaceWeb, API, MCP

Alternatives to VoxLingua107

OpenAI Whisper
OpenAI

The reference open-source multilingual ASR model from OpenAI.

OSS · MIT ★ 98.1k
whisper.cpp
Georgi Gerganov

C/C++ port of Whisper — runs on anything, from a Raspberry Pi to Apple Silicon.

OSS · MIT ★ 48.8k
faster-whisper
SYSTRAN

4× faster than reference Whisper using CTranslate2 — production sweet spot.

OSS · MIT ★ 22.3k

Whipscribe is a managed faster-whisper + whisperX service. If you want transcripts without running infrastructure, paste a URL or drop a file in the form below — you'll have a transcript in seconds.