Datasets | Mozilla Data Collective

All Datasets

Compensated Datasets

API

Contribute Datasets

Loading datasets...

Terms

FAQs

Participation Guidelines

Responsible Disclosure

News and Documentation

Data and Model Competitions

support@mozilladatacollective.com

Terms

FAQs

Participation Guidelines

Responsible Disclosure

News and Documentation

Data and Model Competitions

support@mozilladatacollective.com

Mozilla Data Collective is a mission-locked British company, incubated by Mozilla Foundation, and backed by the not-for-profit Mozilla.org

All Datasets

1,013 Datasets

MDC Community Concierge

VoxForge - German

32 hours (24685 utterances) of read speech of Deutsch (German).

License: GPL-3.0

Locale: de

Task: ASR

Format: WAV, TSV

Size: 11.59 GB

MDC Community Concierge

VoxForge - Greek

4 hours (1377 utterances) of read speech of Ελληνικά (Greek).

License: GPL-3.0

Locale: el

Task: ASR

Format: WAV, TSV

Size: 954.26 MB

MDC Community Concierge

VoxForge - Catalan

39 minutes of Catalan read speech.

License: GPL-3.0

Locale: ca

Task: ASR

Format: WAV, TSV

Size: 143.71 MB

0DIN by Mozilla

Public GenAI Vulnerability Disclosures

Weekly export of 0DIN's public GenAI vulnerability disclosures: jailbreaks, prompt injection, content manipulation, and other model-security findings.

License: CC-BY-4.0

Locale: en

Task: N/A

Format: JSONL

Size: 116.26 KB

MDC Community Concierge

VoxForge - Bulgarian

1 hour of read speech in Bulgarian, collected via the VoxForge project

License: GPL-3.0

Locale: bg

Task: ASR

Format: WAV, TSV

Size: 234.06 MB

MDC Curators

CoVoST 2 German-English

The dataset includes 154856 audio files in German and their translations in English.

License: CC-BY-NC-4.0

Locale: de, en

Task: MT

Format: MP3, TSV

Size: 5.90 GB

CLEAR Global

TWB Parallel Sentence kits - Congo Swahili (25k)

25,305 French–Congo Swahili parallel sentences from CLEAR Global's Gamayun kits, translated by professional translators from Tatoeba source sentences.

License: CC-BY-4.0

Locale: swc

Task: MT

Format: TSV

Size: 2.18 MB

CLEAR Global

TWB Parallel Sentence kits - Swahili (5k)

5,000 English–Swahili parallel sentences from CLEAR Global's Gamayun mini-kit, translated by professional translators from Tatoeba source sentences.

License: CC-BY-4.0

Locale: swa

Task: MT

Format: TSV

Size: 347.61 KB

CLEAR Global

TWB Parallel Sentence kits - Rohingya (5k)

5,000 English–Rohingya parallel sentences from CLEAR Global's Gamayun mini-kit, translated by professional translators from Tatoeba source sentences.

License: CC-BY-4.0

Locale: rhg

Task: MT

Format: TSV

Size: 358.88 KB

CLEAR Global

TWB Parallel Sentence kits - Lingala (5k)

5,000 French–Lingala parallel sentences from CLEAR Global's Gamayun mini-kit, translated by professional translators from Tatoeba source sentences.

License: CC-BY-4.0

Locale: lin

Task: MT

Format: TSV

Size: 494.43 KB

CLEAR Global

TWB Parallel Sentence kits - Nande (15k)

15,000 French–Nande parallel sentences from CLEAR Global's Gamayun kits, translated by professional translators from Tatoeba source sentences.

License: CC-BY-4.0

Locale: nnb

Task: MT

Format: TSV

Size: 1.26 MB

CLEAR Global

TWB Parallel Sentence kits - Tigrinya (5k)

5,000 English–Tigrinya parallel sentences from CLEAR Global's Gamayun mini-kit, translated by professional translators from Tatoeba source sentences.

License: CC-BY-4.0

Locale: tig

Task: MT

Format: TSV

Size: 404.75 KB

CLEAR Global

TWB Parallel Sentence kits - Kanuri (5k)

5,000 English–Kanuri parallel sentences from CLEAR Global's Gamayun mini-kit, translated by professional translators from Tatoeba source sentences.

License: CC-BY-4.0

Locale: kau

Task: MT

Format: TSV

Size: 358.46 KB

MDC Community Concierge

The LJSpeech Dataset

This is a public domain speech dataset consisting of 13,100 short audio clips of a single speaker reading passages from 7 non-fiction books.

License: CC0-1.0

Locale: en

Task: TTS

Format: WAV, TSV

Size: 3.00 GB

CLEAR Global

TWB Parallel Sentence kits - Hausa (30k)

30,000 English–Hausa parallel sentences from CLEAR Global's Gamayun kits, translated by professional translators from Tatoeba source sentences.

License: CC-BY-4.0

Locale: hau

Task: MT

Format: TSV

Size: 1.68 MB

Community

TTS Bugis - Barru Dialect: Language and Identity

This dataset focuses on the Bugis language, specifically the Barru dialect, and encompasses both formal and informal linguistic registers of Bugis.

License: CC-BY-NC-4.0

Locale: bug

Task: TTS

Format: WEBM, TSV

Size: 287.70 MB

Bahasa Bahana

BAHANA-Manggarai TTS

The dataset consists of various short texts written in the Manggarai language and the texts are read and recorded by a native speaker.

License: CC-BY-NC-SA-4.0

Locale: mqy

Task: TTS

Format: WEBM, TSV

Size: 273.32 MB

Bahasa Bahana

BAHANA-Betawi TTS

This dataset consists of various texts involving Indonesian and English code-mixing and read by Betawi speakers of West Java Province, Indonesia.

License: CC-BY-SA-4.0

Locale: bew

Task: TTS

Format: WEBM, TSV

Size: 316.13 MB

Tamazight NLP

Tamazight Open Speech Dataset

This dataset provides a parsed, formatted, and ready-to-use Amazigh Voice Dataset. It contains voice recordings and corresponding text transcripts in Standard Moroccan Amazigh (ⵜⴰⵎⴰⵣⵉⵖⵜ ⵜⴰⵏⴰⵡⴰⵢⵜ ⵜⴰⵎⵓⵔⴰⴽⵓⵛⵜ) intended for training Automatic Speech Recognition (ASR) and Text-to-Speech (TTS) models.

License: Apache-2.0

Locale: zgh

Task: ASR

Format: WAV, JSONL

Size: 459.09 MB

Community

Italian TTS - female voice

Italian, 66yo woman reading "Il fu Mattia Pascal" and "L'Argentina vista com'è", italian classical literature.

License: Apache-2.0

Locale: it

Task: TTS

Format: WEBM, TSV

Size: 680.16 MB

CLEAR Global

Read Speech in Kenyan Swahili (6h)

~6 hours of prompted read speech in Kenyan Swahili from a single anonymous male speaker, with transcriptions drawn from Tatoeba-sourced sentences.

License: CC-BY-NC-4.0

Locale: sw

Task: ASR

Format: WAV, TSV

Size: 1.59 GB

RFE/RL

RFE/RL Belarusian News Text Corpus

Longitudinal Belarusian news corpus from Radio Svaboda (1997-2026) with nearly 339,000 articles and 134M tokens.

License: CC-BY-NC-SA-4.0

Locale: be

Task: NLP

Format: TXT

Size: 486.55 MB

RFE/RL

RFE/RL Afghan Dari News Text Corpus

Longitudinal Afghan Dari news corpus from Radio Azadi (2000-2026) featuring over 212,000 articles and 50M tokens.

License: CC-BY-NC-SA-4.0

Locale: prs

Task: NLP

Format: TXT

Size: 115.70 MB

RFE/RL

RFE/RL Afghan Pashto News Text Corpus

This dataset is a longitudinal news corpus for the Pashto language, sourced from Radio Azadi, from 2000 to 2026, it contains over 204,000 articles (16M tokens).

License: CC-BY-NC-SA-4.0

Locale: ps

Task: NLP

Format: TXT

Size: 117.68 MB