A soundscape composed for the hour you're actually having.

You arrive with a state and a need. You tell Soundscaper what's going on — type it, speak it, pick from the wizard's options. From there a session is composed for you: the length you have, the listening you're doing, and the support area that fits the thing you're trying to ease. Bedtime, focused rest, meditation, work, caring for someone else — each shapes the session differently.

Underneath every session is a stack of techniques the music-therapy evidence base actually supports — receptive music therapy for low mood and for dementia comfort, breath-paced movement at the body's natural cardiorespiratory rhythm, nature recordings for stress recovery, embedded amplitude modulation for attention. Each one is named on the methods page with what the evidence covers and what it doesn't.

Three things hold across every session. Nothing jolts you — transitions are bridges, not cuts. The session walks an arc rather than a flat line: arrive, settle, deepen, sustain, come back. And the words you type in are read by a language model Kenjin runs on its own servers — never sent to a third-party AI service to compose the session.

Every session under four hours walks the same arc — settle, deepen, sustain, integrate, close — linked by short bridges instead of cuts. Past four hours it becomes an overnight: a short gateway, a long sustain scheduling pink-noise pulses into your likely sleep windows, then a handoff to sleep. No cues. The audio tapers to silence and stops.

A support area is a felt need — the state you brought to the audio — not a diagnosis. Each of the eight has its own page: what the literature actually says, what we're using, what we're deliberately not claiming. The wizard composes for whichever combination you arrive with.

The same composition sits very differently in different rooms. The wizard adapts the technique stack to where you are: nature-led for the kitchen, music-led for the desk, the bare composition for advanced practice, the overnight shape for sleep. It suggests a context from your support area; you can override it.

What sits underneath a session can be packaged five ways. Same composition, same specification — what changes is what you actually hear. Some listeners want the technique invisible under nature recordings; some want the music to be the experience; some want it bare. For the nature carriers you can pick which elements you want — water, rain, wind, forest, birdsong, distant thunder, fire crackle, insects — sourced CC0 only. (Spatial rendering for headphone listeners is on the way; it isn't here yet.)

The excerpts are from real rendered sessions, unedited apart from a fade at each end.

Most wellness audio either ignores the evidence or oversells it. We grade every claim into one of three tiers, and the tier is visible wherever the claim appears — the methods page, each support-area page, the science page. Citations link to the underlying papers. You can check.

The words you write or speak about your state are read by a language model Kenjin runs on its own servers — never sent to OpenAI, Google, or any other third-party AI service. Out of that reading comes a structured profile — your support area, length, context, the choices you've made — and that's what composition runs on. Your intake is handled only by our own systems, under our Data Policy. No analytics cookies. No advertising cookies. No behavioural profiling. There's no consent banner because there's nothing to consent to.

The formal posture is at /privacy-policy. The plain-language version is at /data-policy.

Listening is the work. A reflection on the listening is optional — never a streak, never a prompt, never a guilt loop. The journal sits next to every recording in your library, in three short modes you can use or ignore.

Pre-session — a sentence or two about what you're bringing in, captured before you press play. Post-session — what changed, or didn't, after the audio finished. Standalone— a reflection that isn't tied to a session at all.

The journal is yours alone. Nothing leaves your account unless you opt into the Insights surface later. The full walkthrough lives on the mobile app page.

I'm the founder of Soundscaper. This site argues for a category I commercially benefit from, and I've tried to be more sceptical of my own claims than I am of other people's. Where I've slipped, the science page and the boundary section above are where I try to keep myself honest. If you spot a claim here that the evidence doesn't support, email me. That's the credibility loop.

Email: [email protected]