Engineering notes from the
AutoAIClips team.
We write about clip ranking signal design, 9:16 reframe heuristics, and what we keep learning about how TikTok, Reels, and Shorts actually rank short-form content. No guides on “5 tips to go viral” — just the engineering and the data.
Why hook strength beats generic virality scoring
Most AI clippers ship with a "virality score" that conflates a dozen unrelated signals into one number. We unbundle it — and show why hook strength is the only signal that survives algorithm changes.
Read post →
9:16 reframe explained: why speaker tracking beats center crop
Center crop is the lazy default for 16:9 → 9:16 conversion. It cuts off faces in two-person interviews and looks amateur on TikTok. Here is what speaker-aware reframe actually does under the hood.
Read post →
TikTok algorithm 2026: why YouTube watermarks get your clips demoted
TikTok actively demotes content with visible watermarks from other platforms. We trace why this policy exists, how it works in practice, and what clean-export means for your repost strategy.
Read post →