Datasets | Mozilla Data Collective

Bulu (bum) scripted speech dataset: 819 clips from 8 speakers (predominantly female) across 9 recording sessions.

Bulu-ASR-Dataset

License: NOODL-1.0

Locale: bum

Task: ASR

Format: MP3, TSV

Size: 42.11 MB

Field Matters

AIRI AI4TALK Meadow Mari ASR

Meadow Mari ASR data from AIRI AI4TALK with MP3 speech segments and Meadow Mari Cyrillic-orthography transcriptions in CSV.

License: CC-BY-SA-4.0

Locale: mhr

Task: ASR

Format: CSV, MP3

Size: 421.20 MB

Field Matters

AIRI AI4TALK Khakas-Russian Translation

Khakas-to-Russian translation data from AIRI AI4TALK with source text and Russian translations in CSV.

License: CC-BY-SA-4.0

Locale: kjh, ru

Task: MT

Format: CSV

Size: 262.01 KB

2,488 read-speech recordings from a single Dagbani speaker, with MP3 audio and TSV sentence mappings. Suitable for TTS and ASR.

Sample Dagbani-TTS-Dataset

License: NOODL-1.0

Locale: dag

Task: TTS

Format: MP3, TSV

Size: 175.99 MB

Multimodal dataset for ɣɔmáláʔ (Ghomala, Bandjoun variety): IPA-annotated ritual corpus, 369 audio clips (≈17 min), and a Ghomala–French parallel corpus.

Ghomala-MultimodalDataset

License: NOODL-1.0

Locale: bbj

Task: NLP

Format: TSV, MP3

Size: 18.67 MB

1,737 read-speech recordings from a single Afaan Oromo speaker, with MP3 audio and TSV sentence mappings. Suitable for TTS and ASR.

Afaan Oromoo-TTS-Dataset

License: NOODL-1.0

Locale: gax

Task: TTS

Format: MP3, TSV

Size: 258.48 MB

A Telugu text corpus for NLP research, containing diverse written content for language modeling, text analysis, and linguistic studies.

Telugu Text Corpus

License: CC-BY-NC-SA-4.0

Locale: tel

Task: NLP

Format: TXT

Size: 19.87 MB

A curated corpus of Tamil text designed to support natural language processing tasks such as language modeling, text classification, and linguistic analysis.

Tamil Literature Corpus

License: CC-BY-NC-SA-4.0

Locale: tam

Task: NLP

Format: TXT

Size: 41.85 MB

A diverse Hindi text corpus spanning literary works and news article blogs, curated for NLP research and language modeling tasks.

Hindi Literature & News Corpus

License: CC-BY-NC-SA-4.0

Locale: hin

Task: NLP

Format: TXT

Size: 19.52 MB

fastdol.com

Public Company Federal Compliance Records — Q3 2026

Federal workplace enforcement (OSHA, WHD, NLRB, EPA, SAM) joined to SEC parent financials for 103,974 establishments at 1,826 public companies.

License: CC-BY-4.0

Locale: en

Task: OTH

Format: csv

Size: 2.74 MB

Field Matters

AIRI AI4TALK Meadow Mari-Russian Translation

Meadow Mari-to-Russian translation data from AIRI AI4TALK with source text and Russian translations in CSV.

License: CC-BY-SA-4.0

Locale: mhr, ru

Task: MT

Format: CSV

Size: 89.27 KB

The dataset includes audio files in English and their translations in Mongolian.

CoVoST 2 English - Mongolian

License: CC-BY-NC-4.0

Locale: en, mn

Task: MT

Format: MP3, TSV

Size: 12.56 GB

MDC Datasets

Brazilian Sign Language (Libras) Healthcare Corpus (Non Commercial)

Multimodal Libras dataset of ~30h video on health topics, with gloss annotations and Portuguese translations.

License: CC-BY-NC-SA-4.0

Locale: bzs, pt

Task: OTH

Format: MP4, SRT, JSON

Size: 40.31 GB

Ofis publik ar Brezhoneg (OPAB)

TermOfis terminology database Breton-French

TermOfis is a terminology database for the breton language. it contains over 90 000 terms covering around 160 subject fields.

License: CC-BY-NC-SA-4.0

Locale: br, fr

Task: N/A

Format: CSV

Size: 1.69 MB

Amnesia

Buscadores Temáticos de la Plataforma Nacional de Transparencia (PNT) de México

El dataset comprende la totalidad de los buscadores temáticos de la PNT de México, extraídos sistemáticamente por el equipo Amnesia durante 2025

License: CC0-1.0

Locale: es-mx

Task: NLP

Format: CSV

Size: 16.54 GB

fastdol.com

OSHA Severe Injuries by Employer 2015-2025

Employer-level OSHA severe injury records — hospitalizations, amputations, and eye losses for 84,813 U.S. employers with cross-agency enforcement flags.

License: CC-BY-4.0

Locale: en

Task: OTH

Format: CSV

Size: 6.75 MB

A curated Marathi text corpus containing diverse written content for natural language processing, linguistic research, and language modeling tasks.

Marathi Blog & Literature Corpus

License: CC-BY-NC-4.0

Locale: mar

Task: NLP

Format: TXT

Size: 4.20 MB

This dataset comprises example sentences in Gbaya-Lay with French equivalents, audio recordings by a native speaker, and a sentence–audio alignment file.

Gbaya-Lay_ALCAM-MultimodalDataset

License: NOODL-1.0

Locale: gya

Task: NLP

Format: MP3, TSV

Size: 9.23 MB

Community

Mandar Balanipa Speech Corpus

This Mandar language dataset is text and audio data paired for languange documentation and linguistic research in technology.

License: CC-BY-NC-4.0

Locale: mdr

Task: TTS

Format: WEBM, TSV

Size: 281.70 MB

Community

TTS Javanese-Gresik Dialect

This dataset consists of speech audio clips paired with text in the Javanese language with Gresik Dialect used in daily activities.

License: CC-BY-NC-SA-4.0

Locale: jav

Task: TTS

Format: WEBM, TSV

Size: 282.84 MB

The dataset includes audio files in English and their translations in Persian.

CoVoST 2 English - Persian

License: CC-BY-NC-4.0

Locale: en, fa

Task: MT

Format: MP3, TSV

Size: 12.56 GB

Ofis publik ar Brezhoneg (OPAB)

KerOfis toponymic database Breton-French

KerOfis is a database of toponyms and provide the correct forms of Breton place names. The database also contains the Breton forms of foreign place names.

License: ODbL-1.0

Locale: br, fr

Task: N/A

Format: CSV

Size: 1.09 MB

The dataset includes audio files in English and their translations in Turkish.

CoVoST 2 English - Turkish

License: CC-BY-NC-4.0

Locale: en, tr

Task: MT

Format: MP3, TSV

Size: 12.56 GB