Sound-based wellness audio, with the working shown.

Soundscaper composes personalised sound-based sessions for non-clinical wellness use. Every session opens with a headphone induction phase drawing on the Monroe Institute / Hemi-Sync lineage, transitions into a recipe-driven body, and closes with either a return to alertness or a sleep handoff.

For clinicians and researchers: the composition pipeline is documented end to end at /approach, the evidence base for each technique is graded and cited at /science, and the safety boundary — including the wellness-vs-clinical-treatment line — sits at /safety. We don't hide the working.

Every claim is graded into one of three tiers. Music-based interventions in dementia and music therapy for depressive symptoms sit at Established. Binaural-embedded soundscapes for state anxiety, and musical amplitude-modulation for attention, sit at Emerging — with mechanism caveats stated. The Monroe Institute / Hemi-Sync lineage is treated as Speculativein respect of efficacy; it's design lineage, not proof basis.

The science page lists each citation with a publisher link. If a claim on this site doesn't stand up to your reading of the literature, write to us — we'd rather correct it than defend it.

Soundscaper isn't a medical device. It isn't cleared by the FDA, MHRA, or any regulatory body. It doesn't treat anxiety, depression, insomnia, PTSD, ADHD, or any clinical condition. The product is wellness-adjacent listening, for self-directed and carer-mediated use.

Crisis-indicator language at intake routes the user to UK helpline resources (Samaritans 116 123, NHS 111, Mind, Find a Helpline) rather than generating a session. SAFETY_FAIL_OPEN is configured false in production: when the safety classifier is unavailable, the system blocks rather than proceeds. The audio engine holds conservative loudness ceilings (−16 LUFS general, −20 LUFS sleep) and a −1 dBFS true-peak ceiling.

The collaborations that interest us most:

• Care-home and dementia-comfort pilots.The MIDDEL trial established a serious evidence base for music-based intervention in dementia. We'd like to learn how a generated, individualised composition compares to clinician-curated playlists in real care settings.
• Convalescent-care and post-operative anxiety.The 2025 perioperative binaural-beat meta-analysis (~SMD −1.38) is striking. We'd like to support studies that test whether a generated, intake-personalised version performs comparably to fixed binaural recordings.
• Methods reviews.If you're writing about generative wellness audio, sonic personalisation, or the regulatory boundary for sound-based interventions, we'll share our working — composition manifests, intent profile schemas, evidence-tier grading methodology — in full, under whatever terms suit your work.

We aren't in a position to make clinical-grade efficacy claims. We are in a position to support honest research on what individualised generative audio actually does — so the category recovers some of the ground that wellness over-claim has lost it.

There's an early-access programme for clinical and care contexts — clinical advisors, NHS-pathway partners, care-home procurement, hospital wellness programmes, academic researchers. Substantial discount; structured feedback expected. Write to us with a brief description of your context.