01 / THE CRAFT
A seven-dimension assessment.
A craft grade out of 100, reported separately from the audience prediction, so you can distinguish filmmaking execution from audience appeal.
Free private beta · For filmmakers
Test screen your cut.
Tonight. Privately.
AI that watches your entire film, grades its craft across seven dimensions, projects its IMDb score, and gives you timecoded recommendations for your next cut.
For feature-length narrative films, from rough cut to locked picture.
See what you’ll receiveBuilt by iScreeningRoom.
More than ten years of real audience test screenings.
What you get
A complete read of your film.
Explore three views from the full report, then see everything your evaluation includes below.
Tap a preview to enlarge.
Example report views. Scores and findings shown belong to the example films.
Inside your full report
02 / THE AUDIENCE
Predicted reception.
Audience reaction, a predicted reaction distribution, and an IMDb landing zone with a center estimate and range.
03 / THE TIMELINE
Engagement across the entire cut.
A minute-by-minute read of where your film holds attention or loses it, supported by specific, timecoded moments.
04 / THE FINDINGS
What works. What to change.
Detailed findings, an executive synthesis, and recommended changes grounded in what actually happens on screen.
05 / THE POSITIONING
Your film in the market.
How the film plays against its positioning, comparable films with verified ratings and domestic grosses, and an optional key-art assessment if you upload a poster.
06 / THE CONVERSATION
Work through it with Felix.
Ask follow-up questions about the report, examine individual scenes, and discuss how to approach changes to your cut.
The full report also includes methodology and evaluation details, so you can understand how the assessment was made.
Made for the work before release
Get outside perspective.
Keep the audience out.
Evaluate your cut without recruiting viewers or putting screeners into circulation. Processing uses private cloud storage and paid AI APIs.
- Your film is not used to train general AI models.
- Source uploads are deleted after a successful analysis.
- You keep ownership of your film.
A published accuracy record
0.235
Average error against real human panels
on the panel’s 1–5 scale.
Measured on 12 films held out from training and tuning, as of October 2026. This is an average across those films, not an IMDb error figure or a guarantee for any one film.
See the method and limitationsBefore you apply.
What does the beta cost?
Screenings are complimentary for invited filmmakers during the private beta. Access is limited to a small cohort.
Does my film need to be finished?
No. You can evaluate a feature-length narrative film from a watchable rough cut through locked picture. The analysis assesses what is on screen in the cut you upload.
Is the IMDb score guaranteed?
No. It is a projection with a range. The public validation record measures two links: AI predictions against human panels, and panel scores against public IMDb ratings. Neither establishes a promise of an exact IMDb score.
Read the validation record ↗Your film. Your next move.
Find out what’s working.
While you can still change it.
Apply for the free beta Complimentary access for invited filmmakers.