For AI labs

All modalities Voice Robotics Expert evaluation Multimodal & text

For partners

Join the network How partners earn Apply

Live

VoiceTibetan Ü-Tsang pilot launched — Dharamshala RoboticsRobotics teleop cohorts expanding across SE Asia Expert EvalExpert evaluator network expanding across domains VoiceSardinian pilot kicked off via Cagliari + Sassari studios MultimodalCross-modality embodied-AI dataset scoping with frontier lab VoiceCebuano production cohort scaling

FOR AI LABS

Data your existing vendors can't quote on.

Underrepresented-language voice. Robotics demonstrations from regions your team can't recruit in. Credentialed expert evaluation. Multimodal RLHF traces. Quartz Labs sources, annotates, and ships across modalities your default vendors won't cover.

Request a sample bundle →

BUILT BY OPERATORS AND RESEARCHERS FROM

Modality 01 — Voice

Voice.

Multispeaker conversational audio with full human-generated transcripts in underrepresented languages worldwide, with hundreds more available on request. Dialect annotation, speaker diarization, demographic balance on request.

Native-speaker sourcing through universities, recording studios, language NGOs, and home-region and diaspora partner networks worldwide.

Pilot scope: scoped per engagement. Production scope: scaled per language on demand.

Sample language coverage

Language	Region	Family	Sourcing	Status	Tier
Tagalog	Luzon, PH	Austronesian	Studio	LIVE	T1
Cebuano	Visayas, PH	Austronesian	Studio	LIVE	T1
Hiligaynon	Western Visayas, PH	Austronesian	Studio	LIVE	T2
Ilocano	Northern Luzon, PH	Austronesian	Studio	LIVE	T2
Hmong (Daw)	Diaspora · MN, US	Hmong-Mien	Diaspora	LIVE	T2
Madurese	East Java, ID	Austronesian	University	LIVE	T2
Tamazight	Atlas, MA	Afro-Asiatic	NGO	LIVE	T2
Kabyle	Kabylia, DZ	Afro-Asiatic	NGO	LIVE	T2
Sepedi	Limpopo, ZA	Bantu	University	LIVE	T2
Guarani	Asunción, PY	Tupian	University	LIVE	T2
Antillean Creole	Guadeloupe, FR	French Creole	NGO	LIVE	T2
Jamaican Creole	Kingston, JM	English Creole	Studio	LIVE	T2
Sardinian	Cagliari · Sassari, IT	Romance	University	LIVE	T2
Luxembourgish	Luxembourg	Germanic	University	LIVE	T2
Bouyei	Guizhou, CN	Tai-Kadai	University	PILOT	NEXT
Rangpuri	North Bengal, BD/IN	Indo-Aryan	NGO	PILOT	T3
Bhili	Gujarat · MP, IN	Indo-Aryan	NGO	PILOT	T3
Tibetan (Ü-Tsang)	Dharamshala, IN	Sino-Tibetan	Diaspora	PILOT	T3
Balochi (Makrani)	Diaspora · Karachi · Muscat	Indo-Iranian	Diaspora	PILOT	T3
Turkmen	Diaspora · TR · DE	Turkic	Diaspora	PILOT	T3
Uyghur	Diaspora · KZ · DC, US	Turkic	Diaspora	PILOT	T3

Modality 02 — Robotics

Robotics.

Teleoperation demonstration data for foundation-model robotics teams. Sourced through our SE Asia operator network — Hanoi and Jakarta cohorts of trained teleoperators with standard manipulation and locomotion setups.

Per-task structured demonstrations with full state, action, and sensory logging. Operator demographic data, task success metadata, and provenance per session.

Pilot scope: 200 hour engagement. Production scope: scale by cohort expansion.

Modality 03 — Expert evaluation

Expert evaluation.

Credentialed domain experts evaluating frontier model outputs in high-stakes domains. Medical (MD, PA, NP), legal (JD, bar-admitted), financial (CFA, CPA, regulatory), and scientific (PhD with current publication record).

Per-query structured evaluation with rubric, reasoning trace, and audit log. All evaluators under NDA. Provenance per evaluation verifiable on request.

Pilot scope: scoped per engagement. Production scope: standing evaluator networks per domain, scaling on demand.

Modality 04 — Multimodal & text

Multimodal & text.

RLHF, instruction tuning, multilingual classification, code corpora, vision-language pairs, cross-modality datasets. Bespoke pipelines built around your spec and schema — not an off-the-shelf product.

Same network applies. Same operational layer: specify, source, annotate, ship, with consent log per row.

Scope: per engagement. Quoted within 48 hours of the first scoping call.

Deliverable

An abstracted voice sample.

Multitrack audio, word-level timestamps, speaker diarization, demographic metadata, and a signed consent log. Same operational layer in robotics, eval, and multimodal — different schemas.

Sample bundle

Consent IDQLB-VOX-2026-04417

LanguageCebuano (ceb)

RegionCebu City, PH

Speakers2 · F/M

Duration00:00:12.4

Sample rate48 kHz · 16-bit

00:0000:0200:0400:0600:0800:1000:12

00.42SPK_01Maayong buntag. Asa man ta moadto karong adlawa, sa merkado o sa playa?
03.18SPK_02Adto ta sa merkado sa, dayon sa hapon sa playa. Init kaayo karon.
06.91SPK_01Sige. Magpalit pud ko og isda nga bag-o, para mag-sugba ta sa gabii.
10.04SPK_02Hala, sulayi ang tindahan ni Nang Lita. Pinakabaratuhon didto.

The operational layer

Request, source, deliver.

Per-row signed consent. Per-batch sample QA before scale. Daily drops or batched delivery to your bucket on your schema.

01 — Request

Your spec.

Language list, hour count, demographic targets, domain constraints, schema, rights model. We work from a spec you write or one we draft together on the first call.

Pilot quote

Within 48 hours. Indicative pricing published.

→

02 — Sourcing

Our network.

Universities, in-country studios, diaspora cohorts, operator networks, credentialed experts, dataset partners. Native-speaker or domain-expert sourcing per modality.

Quality bar

Acoustic, annotation, or domain QA at sample level before scale.

→

03 — Deliverable

Your bucket.

S3, GCS, or SFTP. Multitrack WAV plus JSON sidecars for voice. Per-task logs for robotics. Structured rubrics for eval. Daily drops or batched.

Consent log

Per-row signed release. Provenance verifiable on request.

How we're different

The default vendor wasn't built for frontier data.

Off-the-shelf vendors are built for languages, modalities, and timelines the last AI cycle paid for. We're built for the cases the next one requires.

Dimension	Off-shelf vendor		Quartz Labs
Language coverage	20 major world languages	→	Underrepresented languages live, hundreds more quotable
Modalities	Voice or text, rarely both	→	Voice + robotics + expert eval + multimodal
Ethical sourcing	Mass-scraped, opaque provenance	→	Per-row signed consent log, verifiable
Pilot to quote	4-week scoping minimum	→	48-hour quote, 14-day kickoff
Commitment	6-month minimum, exclusivity	→	Two-week off-ramp, no exclusivity

Work we've shipped.

One live case so far. New cases land as each engagement matures past production milestones.

Case 01 · Voice · 2026 · Frontier voice AI lab

Seven languages, intro to first shipment in eighteen days.

The lab needed multispeaker conversational audio plus human transcripts in seven Southeast Asian and minority languages. We delivered all seven.

Modality: Voice · Status: Production

Read the full case →

Get in touch

Tell us what you're training.

Book a 30-minute scoping call. We map your modality, scope, languages, and timeline — and quote indicative pricing within 48 hours of the intro.