Skylight’s Curator-Led Video App Builds Custom Feeds

Skylight redirects short-form video to human curators

Skylight’s version 2.0 is out and it takes a different route from TikTok-style feeds: instead of forcing a single algorithmic main stream, the app centers human curators who post and repost videos to build custom feeds others can subscribe to. The result is a flatter, community-driven way to discover short-form content.

Built on Bluesky and the AT Protocol, Skylight first launched in April and has since reached roughly 240,000 downloads and hosted about 100,000 uploaded videos. Early investors include Mark Cuban, and the app leverages Bluesky credentials to tap into a wider open-social ecosystem of more than 38 million users.

The v2 update focuses on curation tools. Users can become curators by posting and reposting videos, or by importing custom feeds via third-party builders. Skylight plans to integrate Graze so custom feeds can be created inside the app. To browse curated channels you swipe left from the main page and tap the plus (+) to add feeds you like.

Skylight also added live-streaming functionality through a partnership with Streamplace, avoiding a ground-up build. Other UX upgrades include a redesigned full-screen player and a split-screen mode so you can watch while browsing—small details that make curator discovery feel more natural.

• Human curation offers clearer intent and context than opaque recommendation algorithms.
• Curated feeds can reduce the noise from low-quality AI-generated content flooding algorithmic timelines.
• Open protocols like AT Protocol enable easier account authentication and cross-app portability, accelerating growth for new apps.

Skylight is tiny but noisy. Co-founders Tori White (CEO) and Reed Harmeyer (CTO) built much of the product as a team of two and drove awareness through TikTok updates—White’s account has become a primary acquisition channel, producing spikes of installs when she posts progress clips.

Why does this matter beyond a new app launch? Many users report fatigue with opaque algorithms and a growing distrust of feeds saturated by synthetic content. Skylight’s model answers a simple question: what if discovery were driven by trusted curators instead of a black-box ranking system? For niche communities, brands, and creators, curated channels act like independent magazines or playlists—focused, contextual, and human.

The open-social stack matters here. Because Skylight runs on AT Protocol, it can authenticate Bluesky users and interoperate with other apps built on the same standards. That portability lowers friction for early adoption and opens possibilities for composable social features across services.

For product and policy teams this trend raises practical questions: how do you measure the value of curator trust versus algorithmic reach? How should moderation, monetization, and discovery tools change when feeds are driven by individuals and small communities? Where do brands place sponsored content without eroding curator credibility?

QuarkyByte’s approach to these shifts is investigative and operational: we map technical integrations to business outcomes, model moderation and monetization trade-offs, and run experiments that quantify engagement and retention from curator-led discovery. Organizations exploring curator-first strategies can prototype feed architectures, test trust metrics, and design migration paths that preserve user identity across open social protocols.

Skylight’s v2 is a reminder that the social feed isn’t settled: product choices about who curates content shape what communities form, what creators succeed, and how trusted attention is allocated. Expect other experiments that blend human curation with open protocols—and watch how businesses adapt when audiences prefer editors over algorithms.