Hackathon prototype to production beta

Every language deserves paid work, not extraction.

VoxVox is a marketplace for under-resourced languages where speakers get paid to translate, narrate, transcribe, review, and context-check work with consent, provenance, and cross-border payout rails built in.

I speak an under-resourced language I need language work or data Open the GitHub draft

61 prototype jobs in the local draft database

23 language codes already represented in demo work

27 audio snippets proving voice-first UX direction

Interledger payment architecture already wired into the draft platform

This page is built from the first product repo, the VoxVox pitch materials, and the local prototype database supplied alongside the draft.

Consent-aware by default Text, audio, video, and image deliverables Built for web and WhatsApp-style onboarding Designed around languages with low written-language access

Ailine with her grandmother, whose language story shaped the project.

Founder story

The product starts with a family language that almost disappeared from daily life.

VoxVox is not framed as language nostalgia. It is infrastructure for people whose language knowledge has value but whose market access, payment access, and legal leverage are thin or missing.

The strongest narrative in the source material begins with Ailine's grandmother, an Otomi speaker, and turns quickly into the commercial gap: speakers still have the words, buyers still need the work, but there is no consent-first market that connects them cleanly.

"My grandmother spoke Otomi. VoxVox is for the people who still have the words, and a reason to say them out loud."

Why now

The market gap is cultural, operational, and economic at the same time.

7,170

living languages

The product thesis is built on a language landscape far wider than the small set most AI systems are trained to support well.

44%

already endangered

Language loss is not abstract. It means the next generation may never be paid for knowledge that still exists today.

1 / 3 mo.

languages still disappear

VoxVox treats translation, narration, and documentation as paid labor before the knowledge leaves the room forever.

$73B

translation market by 2028

The buyer demand is already real. The missing piece is a trustworthy path to the people who can actually do the work.

How it works

A marketplace loop built for speakers, buyers, and downstream trust.

Speaker entry

Join by web or messaging-first flow

Source materials describe onboarding that captures language, dialect, region, and modality even when written-language comfort is limited.

Buyer brief

Post real work, not abstract language support

The draft database already includes community notices, worker-rights copy, health messages, cultural art prompts, and oral-history style tasks.

Delivery

Text, audio, video, and image in one system

The product model is multimodal by design. A speaker can translate, narrate, record, subtitle, review, or submit context notes in the same workflow.

Trust + payout

Consent, provenance, royalties, payment

VoxVox positions payment and permission as first-class product layers, not add-ons added after the data has already been taken.

What makes VoxVox different

The most interesting part is not translation. It is the infrastructure wrapped around it.

Self-healing UI

Missing strings become paid work

Each new language creates new localization jobs inside the product itself. The interface gets better because speakers are paid to teach it how to grow.

Consent and provenance

Every submission can carry terms with it

The deck materials consistently position permission, reuse settings, and auditable lineage as premium product features buyers will pay for.

Royalties + payout rails

Speakers should not be paid only once

The thesis extends beyond task payment. Speakers can be compensated again when their work is relicensed downstream into higher-value datasets.

Prototype proof

The draft already shows a coherent product, not just a mission statement.

VoxVox prototype screen showing open translation jobs, budgets, and deliverable filters. — Find work by language, modality, and pay rate. The draft database shows recruiting, selecting, and draft states across many task types.

VoxVox prototype screen showing text translation, audio narration, and recording workflow for a Quechua health page. — Work delivery is multimodal: text, audio, and video can live together in one submission flow.

VoxVox prototype screen showing a missing UI translation prompt that turns product localization into paid work. — The self-healing UI turns untranslated product strings into jobs that speakers can take directly from inside the platform.

Draft database

submissions already modeled

The local SQLite file includes job applications, job submissions, speaker profiles, and payment-related records rather than only marketing copy.

Voice-first UX

UI strings tracked for audio support

The draft stores translatable interface elements and audio snippets for core navigation and form actions, which is central to the accessibility story.

Payment wiring

pending payment records in the draft

Even at hackathon stage, the system already includes data structures for escrow-style payment continuation, finish steps, and quote handling.

Source set used for this page: the public GitHub draft, the VoxVox pitch-deck materials in this workspace, and the local `db.sqlite3` prototype data supplied with the request.

Who it serves

Different buyers, same core need: trustworthy language work from the people who actually hold it.

Speakers

Earn from knowledge that already exists in the community

Translation, narration, review, subtitling, cultural context, and community-facing content work.
Flexible across dialects, modalities, and varying literacy comfort.
Potential for both task payment and downstream licensing participation.

Buyers

Commission language work without pretending consent can be an afterthought

AI labs, universities, governments, NGOs, publishers, and cultural organizations.
Health, civic, legal, educational, and cultural content use cases already appear in the draft data.
One system for text, audio, image, video, and provenance expectations.

Partners

Add onboarding, funding, or field access where the platform alone is not enough

Language networks, foundations, field linguists, and regional institutions.
Good fit for sponsored pilots, grant-backed language expansion, and buyer-side onboarding.
Especially strong where communities want a counterparty and contract, not just research extraction.

Next step

If you want to fund a pilot, source real language work, or bring speakers into the network, this is the right moment.

The draft product is already specific enough to discuss real use cases. The next stage is hardening the platform, validating pilots, and choosing the first production partnerships well.

Speaker interest Buyer or researcher Investor or partner Review the draft repo