Skylight

TikTok’s real moat was never the algorithm. It was your followers. Every creator who has built an audience on a centralized platform knows the arrangement. You supply the labor and the platform owns the relationship. When the rules change, the payouts shrink, or the ownership changes hands, there is no exit door. You start over.

Skylight is a direct challenge to that arrangement. It is a short-form vertical video app built on the AT Protocol, the open, decentralized network that also powers Bluesky. It comes from Skylight Social, PBC, a Public Benefit Corporation founded by CEO Tori White and CTO Reed Harmeyer. The app is published under the developer handle rcharmeyer, and Mark Cuban is among its backers. It launched on January 21, 2025 with a pitch that sounds almost naive in its simplicity: you own your content, your followers, and your connections, and you can take them with you if you leave. It is free to download and carries a 17+ content rating.

The timing helped. Skylight’s user base grew quickly after TikTok’s US ownership changed, as anxious creators went looking for somewhere less contingent to put their work. That is a respectable start for a young app, and a rounding error next to TikTok’s audience. Skylight’s growth argument rests on a larger figure: the tens of millions of accounts across the AT Protocol network, every one of which can sign in to Skylight immediately.

The Public Benefit Corporation structure matters here too. It legally obliges the company to weigh user wellbeing alongside profit. That is a meaningful signal in a category built on compulsive engagement, though it should be treated as a promise rather than a proof. A charter is not a product decision, and Skylight will be judged on what it ships.

• AT Protocol Portability: Your identity, social graph, and videos live on an open network. If Skylight disappoints you, you can move your account and audience to another AT Protocol service without starting from zero.
• Instant Bluesky Onboarding: Existing Bluesky accounts work immediately, and the app automatically follows the creators you already follow there. Cold-start friction is close to zero.
• Cross-App Visibility: Videos posted on Skylight can be viewed, liked, and replied to from Bluesky and other AT Protocol clients, so your audience isn’t limited to people who installed this one app.
• User-Controlled Feeds: You decide what fills your feed. Options include a following feed, custom community-built feeds, and a recommendation algorithm that users report adapts quickly.
• Built-In Video Editor: Record and edit vertical video in the app, then publish without a round trip through a third-party editing tool.
• Standard Social Mechanics: Profiles, follows, likes, comments, and shares all behave the way anyone who has opened a short-form video app would expect.

The Portability Pitch Is the Product

Strip away the protocol jargon and Skylight’s real innovation is a credible exit. On TikTok, Instagram Reels, or YouTube Shorts, your follower count is a number on someone else’s server. On Skylight, the follower relationship is recorded on the open AT Protocol network, which means a different app can read it. That changes the power dynamic between platform and creator in a way no feature update ever could.

The practical payoff shows up on day one. Most challenger social apps die in the cold-start phase. You download them, find an empty feed and no friends, and delete them within a week. Skylight skips that phase for anyone already on Bluesky. Sign in, and the app pulls in the creators you already follow. Your Skylight posts also surface in Bluesky timelines, so the people who see your first video aren’t limited to Skylight’s install base. That is smart distribution engineering. Skylight inherits an audience instead of having to build one from scratch.

The caveat is cultural. Bluesky’s user base skews text-first: journalists, academics, and the politically engaged. That is not the same crowd that made TikTok a phenomenon. Skylight can borrow Bluesky’s graph, but it cannot borrow TikTok’s creative culture. It has to grow its own, and that is slow work no protocol can shortcut.

Feeds, Choice, and the Cost of Control

Skylight’s feed philosophy is its second big idea. Instead of a single opaque “For You” page, you choose your sources: the people you follow, custom feeds, or the recommendation engine. Users report that the algorithm learns quickly, which matters enormously. Short-form video depends on recommendations landing within the first few swipes, and a sluggish engine is a fatal flaw. On paper, Skylight’s design clears that bar.

The problem is the switching. The most common complaint in user reviews is that moving between feed modes is confusing. It’s an instructive failure. Choice is only empowering when the interface makes it legible, and an unclear toggle turns user control into user disorientation. TikTok’s single-feed design is a deliberate trade, sacrificing agency for zero cognitive load, and it works. Skylight is right to reject that trade. It still needs to make the alternative feel just as effortless. Right now it doesn’t.

Creation Tools

The built-in editor covers the fundamentals: recording, trimming, and preparing vertical clips for publishing. That is the minimum viable creative stack for the category, and Skylight meets it. Nothing in the user feedback suggests the editor rivals TikTok’s mature toolkit or CapCut’s depth, and serious creators will likely keep editing elsewhere and uploading the finished files.

That workflow makes upload reliability critical, and it is exactly where Skylight stumbles. Users report bugs when uploading and playing videos. For a video-first app, that is not a minor issue. It is a failure of the core function. A creator who loses an upload once will forgive it. A creator who loses uploads repeatedly will go back to TikTok, portability or not.

Friction Points

The remaining complaints form a familiar pattern for a young app. Users report font-size accessibility problems, a real failure in a product that wants a broad audience. Users also report a search function that drops them out of their results, so tapping into a video and backing out loses their place. That is a small bug with an outsized cost, because search is how users find new creators on a platform still building its catalog.

None of these problems is architectural. They are polish problems, the kind a focused team fixes over a few release cycles. But they add up. The consensus in user reviews is that Skylight “shows promise but is still maturing,” and that is accurate. The ideas are finished. The execution isn’t.

The PBC Question

Skylight’s Public Benefit Corporation status deserves a skeptical look. Legally, it obliges the company to consider user wellbeing alongside profit. In practice, PBC obligations are loosely enforced, and the real test arrives with monetization, when growth targets meet engagement metrics. For now, the structure tells you what Skylight intends to be. Whether it stays that way once revenue pressure arrives is the question to watch.

Skylight wisely doesn’t reinvent the visual grammar of short-form video. Full-screen vertical playback, swipe navigation, and overlaid engagement controls are conventions users already understand, and borrowing them lowers the learning curve. That is the right call. Novelty belongs in the architecture, not in where the like button sits.

Reliability is where the app currently loses ground. Short-form video depends on instant, uninterrupted playback. The format’s whole appeal is the frictionless loop of swipe, watch, swipe. Any stutter, failed load, or playback glitch breaks the trance that makes the format work, and user reports of playback bugs point directly at that weakness. The font-size complaints suggest the interface doesn’t yet scale gracefully for users who need larger text, which is a basic accessibility gap that should rank high on the roadmap.

The design is serviceable and familiar. The engine underneath needs tuning.

Skylight ships on iOS and Android, and the touch-first, vertical, thumb-driven interaction model is native to both. This is a phone app in the purest sense. There is no meaningful desktop creation experience to speak of, and there doesn’t need to be one.

On iOS, Skylight holds a 4.4/5 rating from 1,366 reviews. That is a solid score but a small sample, and the written reviews read more cautiously than the number suggests. Users rate the vision highly and file complaints about the execution. The font-size issue is especially notable on iOS, where users expect apps to honor system-wide text sizing.

On Android, the core feature set is the same, though the available user feedback skews heavily toward iOS. Android users should expect the same reported bugs, plus the usual challenge a small team faces in testing across a fragmented range of devices, chipsets, and screen sizes. Video encoding and playback are exactly where that fragmentation tends to surface.

The bigger platform story is the AT Protocol itself. Your Skylight content shows up in Bluesky and other compatible apps, so the “platform” is really the network rather than any single app store listing.

Skylight gets the hardest part right. It identifies the actual problem with centralized short-form video, which is the lock-in rather than the content, and answers it with an architecture that makes leaving possible. That is a structural innovation, not a feature, and no amount of polish at TikTok can replicate it without dismantling TikTok’s business model.

What Skylight hasn’t yet done is make that architecture feel effortless. Video apps live and die on reliability, and upload failures, playback glitches, and a disorienting feed switcher are the kinds of friction that send casual users back to the incumbent. The audience Skylight inherits from Bluesky is a real advantage, but it is also a narrow one. The app needs to attract creators who have never heard of the AT Protocol and never will.

We recommend it now if you are on Bluesky or worried about platform risk. The upside is real and the cost is nothing. If you want a polished TikTok replacement today, check back after a few more release cycles. Skylight has built the right foundation. Now it has to build the house.