One reader can tell you how the script landed.
That is useful coverage—but it is still one judgment. A strong verdict may reveal a real strength. A troubling note may reveal a real problem. Or either one may simply reflect that reader’s taste.
Your Story. Our AI. The Perfect Rewrite.
- Where your screenplay stands. A verdict and a score from five AI readers, and where they agree.
- What’s holding it back. Every scene, sequence and act gets its own reading, so you can see which parts work and which need work.
- Where the rewrite starts. For the weak spots, ways to fix them, with what each gains and what it costs, and a chat to work through them.
Analyze my screenplay →
Try the free sample
First three scenes · no card required
$9.99full screenplay analysis and poster
Table readyour script, read aloud ▶ Listen to Green Book
35,000+analyses since 2023
30-day money-back on your first analysisNever used to train AI, by us or our AI providersPrivate by defaultNo subscription
Green Book2018 · AI verdict
Gemini
Highly Recommend
8.8 / 10
Reads asPrestige
A masterfully paced character study that balances prestigious thematic weight with a propulsive, accessible road-movie engine.
DeepSeek
Recommend
7.8 / 10
Reads asMainstream commercial
The central relationship lands, but the episodic middle lacks accumulating pressure.
Grok
Recommend
7.8 / 10
Reads asPrestige
Central relationship chemistry sustains engagement despite an episodic second act.
Claude
Recommend
7.5 / 10
Reads asPrestige
Distinctive dialogue carries incidents that do not yet build into one rising action.
OpenAI
Recommend
7.3 / 10
Reads asElevated commercial
The middle repeatedly resets, and Shirley’s parallel arc needs fuller causality.
You don’t have to rewrite everything. Find the parts that need it.
A note like “this scene drags” describes a symptom. The Layered Read asks twelve specific questions about what might be causing it, and asks them of every scene, sequence and act, and of the whole script.
It looks at things I never thought were even something to be paid attention to, and now I go back to every scene I thought was brilliant and realize it could still be better.
Layered Read is our patent-pending technology that lets you see your whole screenplay at once, then move into the acts, sequences, and scenes shaping the result.
Explore a screenplay you know
Try this in Green Book The first act is the only one that comes back needing work — Lip’s prejudice gets set up, then set up again. One sequence inside it still holds, though: the negotiation where he’s hired. Scene 10 on its own doesn’t, and scene 11 carries it. Open scene 10, then the sequence above it. The first act is the only one that comes back needing work — Lip’s prejudice gets set up, then set up again. One sequence inside it still holds: 1-3, the negotiation where he’s hired. Scene 10 on its own doesn’t, and scene 11 carries it. Open Act 1, then sequence 1-3, then scene 10.
Try this in The Room Almost nothing here survives at any level. What comes closest is scene 4 — Lisa getting dressed, waiting for Mark — because something in it actually changes. Open it and see what the read still gives it credit for. Almost nothing here survives at any level. What comes closest is scene 4 — Lisa getting dressed, waiting for Mark — because something in it actually changes. Open Act 1, then sequence 1-2, then scene 4, and see what the read still gives it credit for.
Try this in The Matrix Every act and every sequence holds. The scenes that don’t are the ones where someone explains the world, or nobody is under any pressure. Scene 31 is the crew eating synthetic breakfast: competent, funny, and nothing in it is forcing anyone. Open it and argue with that. Every act and every sequence holds. The scenes that don’t are the ones where someone explains the world, or nobody is under any pressure. Scene 31 is the crew eating synthetic breakfast: competent, funny, and nothing in it is forcing anyone — and the sequence around it holds anyway. Open Act 2a, then sequence 2a-6, then scene 31.
See the full analysis of Green Book ↗
See the full analysis of The Room ↗
See the full analysis of The Matrix ↗
No notes on that one yet. Try Triage, Pressure or Development.
The same screenplay, read at four sizes.
Scenes. More than half need something. The ones that do are mostly scenes that only set something up, or move them from one place to the next. The scenes where the two of them actually deal with each other are the ones that hold.
Sequences. Group those same scenes into sequences and most come back Keep — a scene that sags on its own is often being carried by the sequence it sits in.
Acts. Only the first act needs work, and no single scene in it is the problem. The act keeps telling us the same thing about Lip: that he is prejudiced. The lemonade glasses go in the bin — and then it tells us again, and again. He ends the act exactly where he started, only now he has a job.
Script. Keep. The relationship carries what the single scenes don’t.
One weak scene is not the same problem as a weak sequence. That is what the four rows are for.
What Triage measures
Triage is the sorting question — what needs attention and what does not. The other readings in the menu tell you why.
Every part of the script is read three ways. Design: is it built to do something. Execution: is it written well. Read: does it hold you. Each one comes back solid, soft or broken.
The band counts those three. Keep — all three solid. Polish — nothing broken, at least one soft. Rework — one broken. Two lenses low — two or more broken.
So a band tells you how MANY readings came back weak, not how badly one of them did. A Rework part can have a single reading far gone; a Two lenses low part can have two that are only just under.
Not fully read means too little of that part was graded to say either way. It is not a position on the scale, which is why it sits outside the colour run.
Now change the reading above, and the same four rows say something else.
Is anything forcing this to happen now?
Scenes. The strong blocks are the places they stop — the dive bar, the concert hall, the dining room in Birmingham. Trouble is waiting for them at the next venue, so that is where the film has something to push against. On the road it is quieter — except when the two of them bring the argument into the car with them, and those drives hold too.
Sequences. The sequences that come back weakest are the ones where they are getting under way — The Road Trip Begins, Road Trip to Pittsburgh, The Offer and the Pawn. A road movie whose roads carry the least pressure.
Acts. Each act pushes a little harder than the one before it, and the last is the strongest — the film tightens its grip just as it is ending.
Script. Start to finish, the trip always has somewhere to be.
The trip always has somewhere to be. The drives never do — and the drives are most of the film.
A Christmas Homecoming reads low too. That one is meant to.
Does anything actually change here, or does it say the same thing again?
Scenes. The weakest row on the chart. Scene after scene leaves everyone exactly where it found them.
Sequences. Grouped into sequences, most of the film lifts. Act 1’s sequences do not.
Acts. So the first act is the only one that stays red — Lip’s prejudice is set up, then set up again.
Script. As strong as this film gets. Across the whole story he is unmistakably a different man.
The change you felt was real. Almost none of it happened inside a scene.
Nowhere else in the film do the parts and the whole disagree this much.
No notes on that one yet. Try Triage, Consequence or Clarity.
Almost nothing here holds, at any size.
Scenes. Two are not at the bottom — scene 4 and scene 13. Everything else is.
Sequences. A handful step up. The rest sit exactly where their scenes do.
Acts. All four come back at the bottom.
Script. The same.
When a screenplay fails everywhere, the four rows agree — which is worth seeing once, because on most scripts they do not.
Does anything here cause what comes next?
Scenes. Four stand clear of the rest: the morning argument, the seduction, the confession, and the announcement at the party. Those are the only scenes that hand anything to the one after them.
Sequences. Three carry it — Devastating Discovery, Birthday Masquerade, The Celebration and Announcement — against a flat floor everywhere else.
Acts. None of it survives the jump. The acts read as low as the dead sequences, not as high as the live ones.
Script. Bottom. Whatever those four scenes set up, the film never collects on it.
A plot that moves in four jolts with dead air in between, and none of it reaches the acts.
Can you follow what is happening?
Scenes. The one reading with no scene at the floor. Taken one at a time, you can tell who wants what.
Sequences. Better again. A few scenes together still track.
Acts. It halves. The same material, grouped larger, stops being followable.
Script. Bottom, with everything else.
Every scene is followable. The film they add up to is not — and that gap is the whole diagnosis.
Clarity runs high on most screenplays, so the number is not the point here. The shape is: of every produced screenplay in our library, this one falls hardest as the units get bigger. Plan 9 from Outer Space is second. Nothing else comes close.
No notes on that one yet. Try Triage or Pressure.
The same screenplay, read at four sizes — and it holds at every one.
Scenes. Most hold. The ones that don’t are all the same kind of scene — someone stopping to explain the world. Even in a script this strong, exposition is where it sags.
Sequences. Every run of scenes holds, start to finish. Most scripts drop at least one.
Acts. All four hold, as they do in most good scripts.
Script. Keep.
Nothing here needs finding — which is the point of showing it. And it is not strong in one department: switch Colour by to Design, then Execution, then Read, and the picture barely changes. Most screenplays are lopsided across those three.
Its own first note is economy: trim the training sequences. The read does not even count that as a miss.
Now change the reading above, and the same four rows say something else.
Is anything forcing this to happen now?
Scenes. One stretch runs low — the scenes aboard the ship, after Neo takes the pill and before the Oracle.
Sequences. Only one sequence goes with them: 2a-6, the call to the Oracle.
Acts. The act above absorbs it completely and comes back strong.
Script. Strong. The film always has somewhere to be.
The one place this film goes quiet is the one place its enemy cannot reach. Nothing in the Matrix can touch a scene set on the Nebuchadnezzar.
It recovers the moment they plug back in — and it is a dip inside a script whose scenes carry more force than almost any screenplay in the library. That stretch is quiet for this film, not for the field.
Bring this level of detail to your screenplay.
Layered Read is included with every full screenplay analysis.
Your writing has a shape.
The five-reader verdicts show how the complete screenplay lands. A separate craft profile scores every scene across 14 dimensions and ranks the averages against the 250+ produced screenplays in our vault.
Two screenplays each from Sorkin, Nolan, Cameron, and Paul Thomas Anderson, plotted on the same axes, come out as recognizably different shapes. You get your own, read as a whole: which axes are carrying the script, and whether the shape reads as intuitive, conceptual, or integrated.
Higher isn’t better. It’s a different shape. The question is whether it’s the shape you meant.
It goes scene by scene and truly unpacks the script in a way that reveals blind spots I couldn’t see on my own.
In the video: two screenplays each from Sorkin, Nolan, Cameron, and Paul Thomas Anderson, plotted on the same craft axes. Watch the full breakdown ↗
Why we built it.
We started in May 2023, when the AI models we used could read only a few pages at a time. We’ve been inventing ever since.
I’m a developer, and I was working on a screenplay. I wanted feedback I could act on, so I wrote code that read it scene by scene. Then I realised other writers could use it, and it became ScriptReader.ai.
I love movies and story theory, and I’ve been having a field day turning ideas into tools writers can use. I like numbers and graphs too; they help me see a script. If they don’t help you, everything they show is also written out in plain words. And you don’t have to read it all: start with what you’re worried about, and skip the rest.
Most of these tools started with something I needed as a writer, or something a writer asked us for. If you wish it did something it doesn’t, tell us at [email protected].
Lemuel HahamFounder, ScriptReader.ai
Read a complete analysis first.
Three produced screenplays, analyzed in full: the same report you would get for yours. Each one has strengths and weak spots you can check against a film you know.
Want to see more? Browse this year’s Oscar-nominated screenplays, the Script Vault of 250+ produced screenplays, and the Hall of Fame: scripts writers submitted to us. Some were analyzed before newer features such as the Layered Read existed, so they won’t all have them.
Or start with your own script, free. The free sample reads your first three scenes in detail, with a table read of them, and includes the story critique, themes, tropes and world building. The five verdicts, the Layered Read and the narrated video come with a full analysis. The table below shows everything in each option.
Try the free sample →What you get, and what it costs.
No subscription: one analysis at a time, or discounted credit packs. A full analysis usually arrives in 1–2 hours. Open any item to see what it looks like in a finished report.
If a scene is working, I know why. If a scene has problems, I know why and what to do to improve it.
If you can’t make your script at least 30% better with ScriptReader.ai, you’re not doing your job as a writer. In bang for the buck, there’s nothing even close.
| What is included | Free sample$0No card required | Standard$9.99One full screenplay analysis | Deep Focus$15.55For harder story problems |
|---|---|---|---|
| Five independent verdicts and one synthesis | — | ✓ | ✓each reader’s strongest model |
| Layered Read — script, acts, sequences, scenes | — | ✓ | ✓ |
| Story, structure, and craft | Story critique, themes, tropes, world building | ✓ | ✓ |
| Characters, dialogue, and emotion | — | ✓ | ✓ |
| Detailed scene analysis | First three scenes | ✓every scene | ✓every scene |
| Table read — your screenplay performed aloud | First three scenes | First three scenesthe rest with craft credits | ✓the whole script |
| A poster and narrated video | — | ✓ | ✓higher resolution |
| Keep exploring after the report | — | 10 chat messages · 10 craft credits | Unlimited chat · 20 craft credits |
| Try the free sample | Analyze a screenplay — $9.99 | Choose Deep Focus — $15.55 |
Prices exclude VAT/sales tax, which is added at checkout based on your billing address where applicable.
We never train on your script Private by default 30-day money-back guarantee on your first analysis
Writing on a paid assignment uses separate professional pricing, selected at checkout.
Compare pricing and bulk credits →
What it does not replace: ScriptReader.ai provides AI coverage and repeatable comparisons. It does not replace a trusted human reader’s lived response to voice, subtext, or tone.
It can’t replace human readers, but it can give you an overall sense: is your script working? And maybe where a script needs more work.
Questions.
How long does an analysis take?
A full analysis usually arrives in one to two hours.
Which AI models read my script?
Five: Claude, OpenAI, Gemini, Grok and DeepSeek. Each one reads your screenplay separately and gives its own verdict. The synthesis then shows where they agree and where they split. Deep Focus uses each reader’s strongest model.
Is my script used to train AI?
No. We never train on your script, and every AI provider we send it to has API terms that bar training on it. Your analysis is private unless you choose to share it.
What file do I upload?
A PDF of your screenplay.
What’s the difference between Standard and Deep Focus?
Both include the five verdicts, the Layered Read, the scene-by-scene analysis, the poster and the narrated video. Deep Focus uses each reader’s strongest model, reads the whole script aloud, and comes with unlimited chat and 20 craft credits. Standard includes 10 chat messages and 10 craft credits.
What are craft credits?
They pay for work after the report: more table-read scenes (1 credit a scene, or 15 for the rest of the script), and brainstorming and scene work.
Can I analyze a revised draft?
Yes. Each draft is its own analysis, and credit packs make repeat drafts cheaper. Many of our 35,000+ analyses are revisions of the same script.
Does it replace a human reader?
No. It gives you AI coverage and comparisons you can repeat on every draft. It doesn’t replace a trusted human reader’s response to voice, subtext or tone.
What if it isn’t useful?
Your first analysis has a 30-day money-back guarantee.
Pro readers didn’t understand the AI premise nor the twist ending. But this app was able to understand the premise and help me see what the adjustments needed were.
A great tool for identifying weaknesses in structure, pacing and underdeveloped supporting characters.
Killing my darlings is not easy. ScriptReader.ai sharpened the knife and wiped the blade.
What does your screenplay actually need?
Five independent reads, one synthesis, and Layered Read down to the scene — usually within one to two hours. Or start with the free sample: your first three scenes, no card required.
We never train on your script Private by default 30-day money-back guarantee on your first analysis