FLEURS
by Google Research
Few-shot multilingual evaluation across 102 languages — n-way parallel speech.
TL;DR
Few-shot multilingual evaluation across 102 languages — n-way parallel speech.
Best for cross-lingual ASR + LID + speech translation evaluation across 102 languages with parallel text. Pricing: free.
Category
Open source
License
—
Stars
—
Last push
—
Pricing
free
Platforms
HuggingFace
What it is
FLEURS (Few-shot Learning Evaluation of Universal Representations of Speech) is the read-speech sibling of FLoRes-101 — 102 languages × ~12h, fully parallel sentences. The reference multilingual benchmark for Whisper, MMS, USM. License: CC BY 4.0.
Best for: Cross-lingual ASR + LID + speech translation evaluation across 102 languages with parallel text.
Watch out for: CC BY 4.0 · ~12h/language · derived from FLoRes-101 sentences (read speech) · small per-language footprint. Cite: Conneau et al., SLT 2022.
Watch out for: CC BY 4.0 · ~12h/language · derived from FLoRes-101 sentences (read speech) · small per-language footprint. Cite: Conneau et al., SLT 2022.
Install / use
from datasets import load_dataset; ds = load_dataset('google/fleurs', 'en_us')Features
| Speaker diarization | No |
| Word-level timestamps | No |
| Streaming / real-time | No |
| Languages supported | 102 |
| HIPAA eligible | No |
FLEURS vs Whipscribe
| Feature | FLEURS | Whipscribe |
|---|---|---|
| Category | Open source | Transcription APIs |
| Pricing | free | free beta |
| Speaker diarization | No | Yes |
| Word timestamps | No | Yes |
| Streaming | No | No |
| Languages | 102 | 99 |
| Platforms | HuggingFace | Web, API, MCP |
Alternatives to FLEURS
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.