Your script as numbers — screen time, absences, eighths, flags
Screenplay metrics are the underrated half of script analysis. Two spreadsheets come out of the studio — one per character, one per scene — and between them they answer questions no prose report gets to: how long is your antagonist off screen, how many of your scenes are more than 70% dialogue, and which ones have a vehicle in them.
No coverage service hands a writer this, and the second file is not really for the writer at all.
Before you upload anything
Nothing changes until you Merge
Every edit opens an A/B compare — nothing changes your scene until you Merge.
Your script is never used to train AI
Zero Data Retention on every model — your script is never used to train AI. Zero Data Retention means the model providers we send your pages to are contractually barred from storing them: your script is processed to produce your result and then it is gone from their side. Read the policy.
You stay the author
These tools read your script and tell you what they found. Where they suggest a change, it arrives as a proposal beside your original, never in place of it.
Your work stays yours, and you can delete it
Your projects stay in your account for as long as you use it. An account left completely inactive for 2 years is notified before anything is removed, and you can delete a project or your whole account at any time. Full terms in the privacy policy.
What “A/B compare” means in practice
- Character Voice
- “Rewrites only this character's dialogue — action and other characters stay untouched.”
- Alternate Takes
- “Two contrasting rewrites of this scene, compared side-by-side — pick one, or blend the best of both.”
Both quoted from the editor itself. Every suggestion lands next to your original with a Merge button; declining it leaves the scene exactly as you wrote it.
What it catches
- Per character: scenes, dialogue lines, speeches, words, average speech length and monologue count.
- Per character: first page, last page, longest absence in scenes, longest absence in pages, and whether they are in the climax.
- Per scene: act, heading, start page, eighths, dialogue lines, action lines and dialogue percentage.
- Per scene: speakers, INT/EXT, time of day and location.
- Per scene: a Slugline OK check — a free formatting validation, and a real one for a first-time writer.
- Per scene: flags for stunt, vehicle, animal and crowd.
From a real report
The two CSV exports from a real 121-page feature. Column names exactly as exported.
- screenplay_character_metrics.csv
- Scenes · Dialogue Lines · Speeches · Words · Avg Speech Words · First Page · Last Page · Longest Absence (scenes) · Longest Absence (pages) · In Climax · Monologues
- screenplay_scene_metrics.csv
- Act · Heading · Start Page · Eighths · Dialogue Lines · Action Lines · Dialogue % · Speakers · INT/EXT · Time · Location · Slugline OK · Flags (stunt / vehicle / animal / crowd)
- Sample insight
- Protagonist in 30 of 40 scenes, 1,955 words
- Sample insight
- Antagonist enters page 50, then absent 14 scenes / 45 pages
The scene file is line-producer data
Eighths, INT/EXT by time of day, and stunt, vehicle, animal and crowd flags are not story metrics. They are the inputs a 1st AD and a line producer use to turn a script into days and money. Having them fall out of the same upload that produced your coverage is the concrete version of the claim that this is one tool for the whole lifecycle.
It is also the honest reason to run analysis early rather than at the end: the numbers you need to schedule are the numbers you need to diagnose.
Slugline validation, free
Every scene gets a Slugline OK check. For an experienced writer this is a formality; for a first script it catches the formatting mistakes that make a draft look amateur before anyone has read a word of the dialogue.
Four questions these files answer immediately
How long is anyone off screen? Sort the character file by longest absence and the structural problems announce themselves — a character the audience has stopped thinking about cannot carry a reveal in the third act.
Which scenes are actually dialogue scenes? Sort the scene file by dialogue percentage. A run of scenes above 70% in the middle of an act is the most common reason a second act feels static, and it is invisible while reading because each scene individually reads fine.
What does this shoot actually contain? Filter the scene file by flag and you have every stunt, vehicle, animal and crowd scene in the script as a list — which is the conversation to have with a line producer before the budget exists rather than after.
And where are the night exteriors? INT/EXT plus time of day, per scene, is the single most expensive pair of columns in a screenplay.
Exports
Both files export as CSV and open in Excel, Numbers or Sheets. The rest of the analysis exports to PDF, Word and Excel, and revised pages to FDX and Fountain.
What script metrics do not do
- They do not tell you whether the script is good. A perfectly balanced dialogue distribution across a boring script is a boring script with good numbers.
- Eighths here are computed from the page, not measured on a stage. They are an estimate for planning, and a 1st AD will adjust them.
- The flags are detection, not judgement — a scene flagged 'vehicle' may involve a parked car nobody drives.
- They do not replace a script breakdown. They feed one.
Fast 300 · Standard 450 · Premium 800 credits. Every generation shows its credit cost before it runs, and a failed generation refunds automatically.
Frequently asked questions
What is in the character metrics export?
One row per character: scenes, dialogue lines, speeches, words, average speech length, first and last page, longest absence in both scenes and pages, whether they appear in the climax, and monologue count.
What are eighths?
The industry's unit for scene length — a page divided into eight parts, used to estimate how long a scene takes to shoot. The scene metrics file carries eighths per scene alongside INT/EXT, time of day and location, which is what a stripboard is built from.
Can I use this for scheduling?
It is the input to scheduling rather than the schedule itself. The same parsed script feeds the script breakdown and stripboard tools, where eighths and scene flags become shoot days.
Will my script be used to train AI?
No. Zero Data Retention is enabled on every model used; your material is never used to train, fine-tune or improve any foundation model. See section 3.1 of the privacy policy.
What format are the exports?
CSV — screenplay_character_metrics.csv and screenplay_scene_metrics.csv — so they open in Excel, Numbers or Sheets.
What does it cost?
The analysis runs at 300 / 450 / 800 credits for Fast / Standard / Premium, and a failed generation refunds automatically.
Upload a script and see what comes back.
Start free with 2,500 credits — no card needed. Every tool shows its cost before it runs, and a failed generation refunds automatically.
What we're learning, once a week
Real numbers on making films with AI — costs, workflows, and what actually works. No pitch, unsubscribe anytime.