Lens AI

Technology that helps cultural knowledge become more accessible.

Speech and language tools that help families and communities document what they know, in the languages they speak, and keep it readable for the generations after them.

Introduction

What is Lens AI?

Lens AI is Auvra’s AI initiative for cultural documentation. Lens AI transcribes and translates oral histories, tags cultural items and makes archives searchable, and people can always review and edit what it produces.

It works inside the Auvra app, alongside Cultural Preservation and My Vault. Its job is simple: make what people preserve easier to read, understand and find, without taking the record away from the person who shared it.

Shown here: an illustrative example of a Lens AI conversation in the app.

Language understanding

Which languages does Lens AI focus on?

Lens AI focuses on indigenous and underrepresented languages, starting with African languages, English and French. Most speech and language technology is built for a few widely written languages, while much of the world’s cultural knowledge is spoken in languages that technology barely hears.

Lens AI is being developed with speech and language technology for these languages, so that a story told in a language that is rarely written down can be as easy to keep and read as one told in English.

Many of these languages are mostly spoken rather than written, mix with other languages mid-sentence, and carry words that have no direct translation. Lens AI is designed to keep the original words visible next to any translation, so meaning is not quietly flattened.

A man in blue adire playing a talking drum, an instrument that mirrors the tones of spoken Yoruba
Language carries cultureSome meaning lives in tone, rhythm and sound, which is why the original recording is always kept.

Oral documentation

How does AI help preserve oral histories?

AI can turn a spoken recording into a written transcript and a translation, so a story told once in one language can be read, searched and understood by people who were not in the room, while the original voice is kept.

  1. Step 1

    Record

    An elder tells a story in their own language, as a voice note or a longer recording.

  2. Step 2

    Transcribe

    Lens AI drafts a written transcript of what was said, kept next to the original audio.

  3. Step 3

    Translate

    A translation is drafted for relatives who speak another language.

  4. Step 4

    Review

    The family checks names, places and meanings, and corrects anything Lens AI got wrong.

AI-assisted drafts always sit beside the original recording, never in place of it.

Cultural organisation

How does Lens AI organise cultural knowledge?

Lens AI suggests tags for the people, places, languages, practices and occasions in a record, so a family’s or community’s archive can be organised without hours of manual labelling.

Good organisation is what lets someone find a grandmother’s recipe or a village festival recording years later. Tags are suggestions: you accept, edit or remove them, and you can always add the context only you know.

Scan

What does Scan do?

Scan lets you point your phone camera at an object, a fabric, a carving or an instrument to capture a clear record of it. Lens AI suggests context about what it may be and where it may come from, which you can review and add to, so its story is preserved with it.

  • 1PointAim your camera at a mask, a fabric, an instrument or a carving.
  • 2CaptureSave a clear record of the object, with the photo kept as the original.
  • 3Add its storyLens AI suggests context for you to review, then you add what you know and keep it in your Vault.
An elder in a feathered cap at a ceremony

Human context

Does Lens AI replace human knowledge?

No. Lens AI supports people; it does not replace their knowledge or authority. The elders, families and communities who hold cultural knowledge remain its authorities, and everything Lens AI produces can be reviewed and edited by them.

  • Transcripts, translations and tags are drafts until a person accepts them
  • Corrections by the contributor take priority over anything the AI suggested
  • AI never decides whether a cultural record is authentic

Responsible AI

How is Lens AI built responsibly?

Lens AI is built on five commitments: consent, attribution, privacy, cultural context and human oversight. Together they mean that people stay in control of their culture, and the AI works for them.

Questions

Lens AI, answered

See all FAQs

What is Lens AI?

Lens AI is Auvra’s AI initiative for cultural documentation. Lens AI transcribes and translates oral histories, tags cultural items and makes archives searchable, and people can always review and edit what it produces.

Which languages will Lens AI support?

Lens AI starts with African languages, English and French, with a focus on indigenous and underrepresented languages. The list is expected to grow as communities contribute recordings with their consent.

What is Scan?

Scan lets you point your phone camera at an object, a fabric, a carving or an instrument to capture a clear record of it. Lens AI suggests context about what it may be and where it may come from, which you can review and add to, so its story is preserved with it.

Can I edit what Lens AI produces?

Yes. Every transcript, translation and tag is a draft that the person who shared the memory can review and correct. Lens AI supports people; it does not replace their knowledge or authority.

Does Auvra use my content to train Lens AI?

By default, no. Auvra does not use your personal content, including private Vault content, to train or improve Lens AI. Any programme to improve Lens AI would be opt-in, through a separate consent flow. See the AI Policy for details.

Give every voice a lasting record.

Start preserving the stories and languages your family carries today, with Lens AI helping you capture and translate them.