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

[email protected]

Terms

FAQs

Participation Guidelines

Responsible Disclosure

News and Documentation

Data and Model Competitions

[email protected]

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

All Datasets

Not sure where to start?

Describe what you are building and the Data Assistant will find the right datasets for you.

Try the Data Assistant

Filtered by

Search results for “Hazargi Corpus for Speech Recognition”

100 Datasets

Digital Divide Data

ddd-kenya-somali-68hrs-asr-part1

A collection of high-quality, crowdsourced audio recordings and verified transcriptions for Somali language, produced by Digital Divide Data

License: CC-BY-4.0

Locale: som

Task: ASR

Format: WAV, TSV

Size: 7.68 GB

Community

Oro_Word

Afaan Oromoo word-level speech dataset collected to support open-source speech recognition and text-to-speech technology.

License: CC0-1.0

Locale: om

Task: TTS

Format: .WAV, CSV

Size: 1.28 MB

MirasAICompensated

🚧 Noakhali 10 Hours TTS (Male Speaker) 🚧

A high-quality Noakhali TTS dataset with paired speech recordings and text transcriptions for speech synthesis, linguistic research, and language preservation.

License: MirasAI Data License

Locale: oak

Task: TTS

Format: WAV, TSV

Size: 2.87 GB

RFE/RL

RFE/RL Kazakh News Text Corpus

Longitudinal Kazakh news corpus from Radio Azattyq (2003-2026) with over 139,000 articles and 35M tokens.

License: CC-BY-NC-SA-4.0

Locale: kk

Task: NLP

Format: TXT

Size: 126.81 MB

Institute of African Digital Humanities

Bulu-ASR-Dataset

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

License: NOODL-1.0

Locale: bum

Task: ASR

Format: MP3, TSV

Size: 42.11 MB

MDC Community Concierge

VoxForge - Hebrew

32 minutes (307 utterances) of read speech of ‭עברית‬ (Hebrew).

License: GPL-3.0

Locale: he

Task: ASR

Format: WAV, TSV

Size: 126.98 MB

CLEAR Global

Synthetic Text Corpus for African Language ASR

LLM-generated synthetic text in 10 African languages with human linguistic quality evaluations

License: CC-BY-NC-4.0

Locale: bm,ny,ha,kr,luo

Task: NLP

Format: TSV

Size: 746.63 KB

Common Voice

Mozilla Common Voice Spontaneous Speech ASR Shared Task Train/Dev Data

This datasheet is for the bundle of Mozilla Common Voice spontaneous speech datasets to be used in the Shared Task on Spontaneous Speech.

License: CC0-1.0

Locale: mul

Task: ASR

Format: mp3

Size: 4.30 GB

MDC Community Concierge

VoxForge - Turkish

3 hours (1717 utterances) of read speech of Türkçe (Turkish).

License: GPL-3.0

Locale: tr

Task: ASR

Format: WAV, TSV

Size: 692.29 MB

MirasAI

Ormuri Literature corpus by Rozi Burki

Ormuri Literature Corpus is a collection of Ormuri literary texts by Rozi Burki, supporting research and preservation of this under-resourced language.

License: CC-BY-NC-SA-4.0

Locale: oru

Task: NLP

Format: PDF, DOCX

Size: 18.92 MB

Collaborative Action For Research & Development (CARD)

Gawri (گاؤری) Magazine Corpus

The Gawri (گاؤری) Magazine Corpus is a curated collection of monthly Gawri magazine text (~67,724 tokens) that supports research and language technology.

License: CC-BY-NC-4.0

Locale: gwc

Task: NLP

Format: TXT

Size: 146.71 KB

Community

Sindhi Corpus

SdCorpus is a large scale monolingual Sindhi text corpus collected from publicly available web sources for NLP and language model development.

License: CC-BY-NC-4.0

Locale: snd

Task: LM

Format: TXT

Size: 321.00 MB

Common Voice

Common Voice 7.0 - Single Word Target Segment

This dataset contains the numbers 0 to 9 and the words "yes" and "no" in 34 languages. It contains 84 validated hours of speech.

License: CC0-1.0

Locale: mul

Task: ASR

Format: TSV, MP3

Size: 3.51 GB

MirasAI

Wakhi Literature Corpus

A literature corpus of Wakhi language texts, preserving the language’s literary, cultural, and linguistic heritage for research and low-resource NLP.

License: CC-BY-ND-4.0

Locale: wbl

Task: NLP

Format: DOCX, TXT

Size: 74.83 KB

Taruen

Uzbek Proverbs Text Corpus

A corpus of traditional Uzbek proverbs digitized from a 2005 volume, designed for NLP tasks and LLM cultural competence evaluations.

License: CC0-1.0

Locale: uz

Task: NLP

Format: TXT

Size: 142.68 KB

Community

Sindhi Word Segmentation (SdSEG)

A sentence-level annotated dataset for training and evaluating sequence labeling models such as CRF, BiLSTM, and transformer-based architectures.

License: CC-BY-NC-4.0

Locale: snd

Task: NLP

Format: JSON

Size: 1.13 MB

Digital Divide Data

ddd-kenya-somali-68hrs-asr-part3

A collection of high-quality, crowdsourced audio recordings and verified transcriptions for the Somali language, produced by Digital Divide Data.

License: CC-BY-4.0

Locale: som

Task: ASR

Format: WAV, TSV

Size: 1.33 GB

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

Digital Divide Data

ddd-kenya-somali-68hrs-asr-part2

A collection of high-quality, crowdsourced audio recordings and verified transcriptions for the Somali language, produced by Digital Divide Data.

License: CC-BY-4.0

Locale: som

Task: ASR

Format: WAV, TSV

Size: 8.07 GB

MDC Community Concierge

IsiZulu Second Language Learner Speech Corpus

Gold standard recordings from isiZulu teachers and recordings from L2 learners annotated by isiZulu teachers for phonemic and tonal pronunciation errors .

License: CC-BY-SA-4.0

Locale: zu

Task: CALL

Format: WAV, SQLITE

Size: 5.26 GB

TidyVoice2026 Challenge

TidyVoiceX_ASV

This dataset is designed for speaker verification using the Mozilla Common Voice corpus across 40 languages. It includes approximately 5,000 speakers who each have recordings in more than one language. Leveraging this multilingual overlap, we construct the trial pairs to explore cross-lingual variation in the speaker verification task.

License: CC0-1.0

Locale: mul

Task: OTH

Format: WAV

Size: 36.72 GB

Institute of African Digital Humanities

Tupuri-Bango_TTS-Dataset (female voice)

Tupuri-Bango (tui) scripted TTS dataset: 2,034 clips (3h 53m 42s), single female speaker, AGLC orthography. Female-voice complement to the male-voice dataset.

License: NOODL-1.0

Locale: tui

Task: TTS

Format: MP3, TSV

Size: 169.22 MB

RFE/RL

RFE/RL Pashto (Pakistani) News Text Corpus

Longitudinal Pakistani Pashto news corpus from Radio Mashaal (2010-2026) with over 52,000 articles and 14.5M tokens.

License: CC-BY-NC-SA-4.0

Locale: ps

Task: NLP

Format: TXT

Size: 39.26 MB

Forum for Language Initiatives

Khowar Word List

A UTF-8 encoded Khowar lexical corpus containing alphabet definitions and a structured word list for linguistic research and language technology development.

License: CC-BY-NC-4.0

Locale: khw

Task: NLP

Format: TXT

Size: 64.22 KB