Hear it. Read it. Download it.

Radio archive

A cleared excerpt from a station back-catalogue: in-language talk programming with the presenter, the callers and the studio bed intact. What a rights audit produces at the far end.

Files

• ExcerptAwaiting cleared material

  MP3 proxy · masters on execution

• TranscriptAwaiting cleared material

  JSON · timestamped, speaker-attributed

• Signed station licenceUnder NDA

  Scanned instrument · under NDA

This slot is wired and empty. It fills when a cleared tranche of this type lands — we would rather show you a gap than a stand-in, because a buyer who works out that a sample was synthetic stops believing the manifest too.

Delivery specification

Format

WAV, 16-bit PCM, 48 kHz

Channels

Separated where the source allows; mix documented where not

Metadata

Station, broadcast date, language, speaker count, overlap ratio

Rights

Station licence naming AI training, sublicensable onward

Podcast network

Multi-speaker episode audio from a network back-catalogue, delivered per-speaker where the network recorded it that way — which is most of them, and is why podcast archive is worth more per hour than broadcast.

Files

• ExcerptAwaiting cleared material

  MP3 proxy · stems on execution

• TranscriptAwaiting cleared material

  JSON · diarised

• Signed network licenceUnder NDA

  Scanned instrument · under NDA

This slot is wired and empty. It fills when a cleared tranche of this type lands — we would rather show you a gap than a stand-in, because a buyer who works out that a sample was synthetic stops believing the manifest too.

Delivery specification

Format

WAV, 16-bit PCM, 48 kHz

Channels

Per-speaker stems where the source holds them

Metadata

Show, episode, speakers, language, code-switch events

Rights

Network licence naming AI training, sublicensable onward

Conversational capture

Unscripted multi-speaker exchange recorded to brief on lavalier — overlap, interruption and code-switching preserved in the transcript rather than cleaned out of it.

Files

• ExcerptAwaiting cleared material

  MP3 proxy · WAV masters on execution

• TranscriptAwaiting cleared material

  JSON · code-switching preserved

• Consent instrumentUnder NDA

  Scanned instrument · under NDA

This slot is wired and empty. It fills when a cleared tranche of this type lands — we would rather show you a gap than a stand-in, because a buyer who works out that a sample was synthetic stops believing the manifest too.

Delivery specification

Format

WAV, 16-bit PCM, 48 kHz

Channels

Channel-separated per speaker

Metadata

Speaker labels, diarisation, overlap ratio, accent, acoustic conditions

Consent

Signed release plus recorded verbal, per speaker

Film & documentary

Production archive from Nollywood and independent documentary — cleared per asset, with third-party music and licensed insert footage removed before a tranche is offered rather than after a buyer finds it.

Files

• Proxy clipAwaiting cleared material

  1080p proxy · masters on execution

• Shot listAwaiting cleared material

  JSON · scene and shot boundaries

• Chain-of-title packUnder NDA

  Per-asset instruments · under NDA

This slot is wired and empty. It fills when a cleared tranche of this type lands — we would rather show you a gap than a stand-in, because a buyer who works out that a sample was synthetic stops believing the manifest too.

Delivery specification

Container

MP4, H.264 high profile, faststart

Resolution

Source-limited master; 1920×1080 proxy

Metadata

Production, year, language, scene and shot boundaries

Excluded

Third-party music, licensed inserts, unreleased performers

Egocentric capture

First-person video of working environments, with the written context layer that is the actual product — task, intent, objects handled, and the sequence in which something failed and the operator changed method.

Files

• Proxy clipAwaiting cleared material

  1080p proxy · 4K masters on execution

• AnnotationDownload

  JSON · schema v1.0 · complete, no NDA

• Consent instrumentUnder NDA

  Scanned instrument · under NDA

Delivery specification

Format

3840×2160, 30 fps, H.264, no HDR

Angles

Head or chest primary plus static wide, clap-synced

Labels

Task, intent, objects, hands-visible, failure events, step timings

Consent

Signed release plus on-camera verbal, per subject

Document corpora

Scanned pages at delivery standard with their structure captured rather than inferred — regions, reading order, tables, stamps and handwriting flagged. Degradation preserved, because for a document model the degradation is the signal.

Files

• Sample pageAwaiting cleared material

  PNG preview · TIFF master on execution

• Page structureAwaiting cleared material

  JSON · regions and reading order

• Holder agreementUnder NDA

  Scanned instrument · under NDA

This slot is wired and empty. It fills when a cleared tranche of this type lands — we would rather show you a gap than a stand-in, because a buyer who works out that a sample was synthetic stops believing the manifest too.

Delivery specification

Capture

400 dpi colour, deskewed, one page per image

Master

TIFF per page, PDF/A compile per document

Structure

Regions, reading order, tables, stamps, handwriting flags

Redaction

At source, under the holder's direction, before scanning

Structured records

Schema and data dictionary for the record lines. The schema is public because it is what you need to decide whether to start a conversation; the records themselves move only under an agreement with the institution that holds them.

Files

• Schema & data dictionaryDownload

  JSON · field definitions and types

• Transformation log formatDownload

  CSV · one row per operation applied

• RecordsUnder NDA

  Under institutional agreement only

Delivery specification

Formats

Parquet, CSV or JSONL against your stated schema

Ships with

Data dictionary, transformation log, null and outlier report

Personal data

Removed before delivery, under the holder's supervision

Geography

Generalised to an agreed precision per tranche, never dropped