Don't waste a week on the wrong upload.

Test your next video idea before you film. Drop the title and thumbnail, see a HIT / MAYBE / FLOP verdict in 30 seconds, and walk into production with conviction — not the quiet dread that you're about to film a flop.

Free · No credit card · Beta access through July 1

1.6M+
YouTube videos analyzed
700k+
Verified Day-30 outcomes
14
Niches calibrated
+500k/mo
New videos · growing daily
0.77
Calibrated accuracy (Spearman)

Same niche. Very different outcomes.

Two videos about YouTube growth. One will flop, one won't.
Here's what the scores say before either gets recorded.

34 FLOP · Below average
"10 YouTube Tips That Will Grow Your Channel"
12% hit rate in Education × Mid (100K-500K) · n=1,247
Generic phrasing — 11 of top 12 cohort comparables came from channels 2-5× your size
No curiosity gap — similar listicle patterns cluster at p25 VPD in your cohort
82 HIT · Top quartile
"I Posted Every Day for 90 Days (What Nobody Tells You)"
62% hit rate in Education × Mid (100K-500K) · n=1,247
Specific timeframe ("90 Days") — high-VPD pattern, p82 in your cohort
First-person vulnerability hook — top-quartile phrasing for your channel size

From idea to outcome — in one loop.

1

Connect your channel

Sign in with the Google account that owns your YouTube channel, then connect YouTube Studio. Connecting calibrates every verdict to your actual subscriber count and recent average views — sharper than a niche-and-size baseline — and auto-tracks how your real uploads perform.

2

Drop an idea

Working title. Niche. A thumbnail mock if you have one. Three fields. No script, no description, no SEO ritual.

3

See the verdict

HIT, MAYBE, or FLOP in 30 seconds — with the one or two specific reasons. Refine the title, swap the thumbnail, watch the score move live.

4

Publish, then come back

After you film and post, link the video to your idea. hitornot shows predicted vs actual — where the call was right, where it wasn't, what you can trust next time.

Predict. Plan. Learn.

Predict — before you film

Drop a working title and thumbnail. Get a calibrated HIT / MAYBE / FLOP verdict in 30 seconds — backed by a model trained on 1.6M+ real YouTube videos with verified Day-30 outcomes, tuned to your niche and channel size. Every verdict comes with the specific reasons the model moved that way. Not a vibe check.

Plan — every idea in one place

Workspace for in-flight ideas. Refine the title, swap the thumbnail, see the verdict update live as you iterate. Decide which idea to film this week — and which to scrap before you spend a day on the wrong one. No more lost notes app drafts.

Learn — after you publish

Link your published video back to its idea. hitornot shows predicted vs actual side-by-side — where the call landed, where it didn't, and the metrics behind it. Calibrate your own gut faster than another year of guessing alone.

The tools behind the verdict.

A single score is a start. The decision still belongs to you.
These three surfaces are how creators audit the model before they trust it.

Battle Arena

Two titles in mind? Three thumbnails? Drop them into the arena and watch the model score each variant head-to-head — side-by-side, against the same niche cohort. Pick the winner with conviction, save the runners-up to your workspace for another day. No more posting the wrong variant.

X-Ray any video

Paste any YouTube URL — yours or someone else's. See what the model would have predicted before publish, and how the video actually performed against its niche × channel-size cohort. The honest gap, the cohort comparables, and which features the model leaned on. Trust the verdict by auditing it on videos you already know.

Why the model said that

Every verdict opens to a diagnosis panel: the top three features that drove the score, the cohort comparables, and counterfactual rewrites that show exactly how the score moves if you change the title or thumbnail. No black box. The model decides, the panel explains — so the next iteration is informed, not a guess.

What creators are testing right now

Real predictions, anonymous. Updated daily.

Loading top ideas…
Test your idea free →

Free through July 1, 2026.

hitornot is in beta. Every feature unlocked for every signed-in user — predictions, Title Compare, niche comparables, the whole verdict engine. No credit card. No catch.

Common questions

What is HitOrNot?

HitOrNot is the tool you reach for the night before you film. Drop your title and thumbnail, get a HIT / MAYBE / FLOP verdict in 30 seconds — so the doubt you're carrying about the idea gets answered before you spend a week making it. Stop second-guessing. Test before you film.

How many free predictions do I get?

During beta (through July 1, 2026), every signed-in user gets unlimited access — no credit card, no per-month cap. Each prediction gives you a HIT / MAYBE / FLOP verdict, what's hurting the idea, and a realistic view range. Connect YouTube Studio to make predictions reflect your actual channel baseline.

Do I need to connect my YouTube channel?

Yes — predictions require your real channel context (subscriber count, recent average views, niche) to be accurate. Without it, every creator would get the same generic baseline and the verdict would be meaningless. Want to try the engine first without committing? Use Audit any video — paste any YouTube URL and see what the model would have predicted, no sign-up required. When you're ready, sign in with the Google account that owns your channel. We only read public channel info (name, subs) and your video performance metrics; OAuth tokens stay server-side. Disconnect any time in Account.

What's the workspace?

A planning surface for in-flight video ideas. Drop a title and niche, see the verdict live as you refine. Save up to 50 ideas during beta. Mark each as draft / ready to film / filmed / published. Beats keeping ideas in a notes app and forgetting which ones actually scored well.

How does the retrospective work?

After you publish a video, link it back to its workspace idea. hitornot matches it to the original prediction and shows predicted vs actual side-by-side — what the model got right, what it didn't, and your real CTR + retention numbers. First results appear ~24h after Studio sync; the comparison firms up by Day 7.

Is my video idea private?

Yes. Your ideas are only visible to you. The anonymous leaderboard shows scores only — never titles, niches, or any identifying details unless you explicitly choose to share your result.

How accurate is the prediction score?

Accurate enough to catch the wrong idea before you spend a week filming it. The verdict reliably separates ideas that will land from ones that won't, tuned to creators in your niche and at your channel size. No tool can guarantee views — but the worst videos are the ones you make AND no one watches. Skipping those is the win.

How is this different from YouTube SEO tools?

SEO tools analyze keywords after you publish. hitornot scores your idea before you record a single second, then closes the loop with predicted-vs-actual after you publish. It's the difference between a weather forecast and reading yesterday's weather report — one helps you decide, the other just explains what already happened.

Don't film another flop.

30 seconds to sign in. Free during beta. No credit card. Connect your channel and your first prediction is one click away.

Try it free →