Know why your best
posts worked.
Create more of them.

We send your top 10 winners into the prompt as few-shot examples, plus a voice profile pulled from your captions and transcripts: pacing, vocabulary, the exact phrases you keep using. The model is solving "what would this creator say next" not "what's a generic viral hook." If the output ever reads like ChatGPT, we treat that as a bug and reshape the prompt. That work is most of what we do.

Yes. Patterns are relative to your own baseline, not to a million-view threshold. If your channel sits between 2k and 4k views and one post hit 9k, that's a 2-3× lift and we'll tell you which factors moved it. "Winners" are defined inside your distribution. The math doesn't care if you're at 5k or 500k.

This is one of the cases the pattern detector is built for. Three winners in a row narrow the pattern fast: same hook type, same posting hour, same format choice, often a thread of factors you can ride. We mark which of those signals are repeatable vs lucky and which to lock into your next 5 posts before the lift fades.

Only if your winners already are. The generator imitates your existing voice. If you make documentary-style cooking content with slow openings, you'll get more of that, scored higher. The system has no "viral hook" prior; it has your prior. The hooks that come out look like you on a good day, not like someone else's account.

Neither. You give us your public username, that's it. No OAuth, no permissions screen, no password. We pull what's already public via our data providers. The trade-off is we can only see public data, which means no retention curves on Instagram (Meta doesn't expose them) and we tell you that explicitly in the dashboard. No magic, no hidden access.

First headline insight under 10 minutes for most channels. The dashboard fills in progressively: metadata in seconds, post analyses over the next 5-7 minutes, per-post "why it worked" narratives last. You're not staring at a spinner, you're watching it build. If after 15 minutes there's no insight worth keeping, the trial is free to walk away from.

No. Cancel from the customer portal in one click and we drop your channel data within 14 days. We don't email you a retention pitch, we don't escalate, we don't "pause" your subscription on you. You came, you saw, you left. That's a clean ending and the only kind we run.

Not yet. YouTube is on the v2 roadmap because the analytics surface is genuinely different (retention curves, AVD, CTR all matter and are exposed differently than IG/TT). We'd rather ship Instagram and TikTok at a level that earns trust than bolt YouTube on with a worse pipeline. If you're a YouTube-first creator, this isn't your tool right now and we'll tell you that on the signup page.

Patterns are derived from your raw post data and labeled with the exact post IDs they came from. Every claim links back to the posts that produced it, with engagement numbers you can verify against your own analytics. Nothing is invented. If we say "contrarian hooks lift you 3.4×," you can click into the four posts that proved it. Every number on the dashboard is auditable.