Stillwater.

About Stillwater

An editorial sleep channel for connected TV.

Stillwater began with a frustration. The streaming services where people fall asleep are noisy by design. Bright thumbnails, frantic promos, autoplay carousels — every screen optimized for attention when you're trying to lose attention. Even the dedicated sleep apps tend to feel like productivity software in disguise.

We wanted a channel that behaved like a quiet hotel room. Lights low. A window slightly open. Something to look at, something to listen to, and a refusal to escalate. That's Stillwater.

What we make

Fifteen scenes at launch, growing to thirty within ninety days. Each scene is one image — a fireplace, a coastal dusk, a moonlit lake — held for as long as you want it. Each scene has its own ambient track, composed specifically for the visual rather than borrowed from a library. No talking by default. Sleep stories will arrive in a separate category later.

How we make it

Visuals come from human-shot footage in the public domain (mostly Pexels, with attribution preserved). We curate, color-correct, and delivery-encode each scene. Music is composed in Suno V5 under a Pro commercial license, with multiple takes per scene before we choose and master one. Channel artwork uses FAL nano-banana 2 with explicit human direction. We disclose AI assistance in every place the platform requires it.

AI content disclosure

Stillwater uses AI as a production tool with humans in every loop. We disclose this here, in our Roku channel manifest, in the Roku Channel Store submission form, and on every page where the platform asks. We are not hiding it.

What is AI-generated: the original ambient music track that plays alongside each scene is composed using Suno V5 (Pro commercial license, which includes broadcast and streaming rights). Each track is generated from a written prompt by our music director, listened through end-to-end by a human, mastered, and approved before it is paired with a scene.

What is not AI-generated: none of the video content. All visuals are licensed from human creators on Pexels under a commercial-use license, with attribution preserved in our records and inside each episode's metadata. None of the narration, voice, or dialogue (when Sleep Stories launches in Q3 2026) will be AI-generated; we plan to commission human voice-over performers under standard talent agreements.

Channel artwork and marketing imagery use Google Imagen / FAL nano-banana 2 under explicit human prompting and editorial selection. We treat these as illustrations rather than photographs and do not present them as documentary content.

Why we're transparent about this: Roku, the FTC, and several state advertising regulators are tightening rules around synthetic content in 2026. We support those rules. A sleep channel that lies about how it's made cannot ask viewers to relax. We'd rather lead the disclosure than be caught not leading it.

Questions about how a specific scene was made? Write to team@stillwatertv.com and we'll send the source attribution and any AI prompts used in production.

Where we operate from

Stillwater is operated by WYSAMM LLC, an independent media company registered in the Republic of Armenia (TIN 08222398), with its registered office in Yerevan. WYSAMM also operates Adbustr, a transparent supply-side platform for connected-TV inventory, which provides part of Stillwater's ad fill.

Why "Stillwater"

A still body of water doesn't demand attention. It absorbs it. That's the channel.

The team

A small editorial group based between Yerevan and Bali, with engineering, design, and music supervision contributed by independent practitioners. We're intentionally small: the work is curatorial rather than productive.

Roadmap

Contact

For anything not covered on the support page, write to team@stillwatertv.com. We read every email.