Frequently asked questions
What exactly does SCRIPTRISK do?
It reads a screenplay the way a studio clearance department does. Four AI desks — clearance counsel, ratings board, safety underwriter, and territory censors — investigate it concurrently and return a Production Risk Report: severity-ranked findings, each with a citation, a concrete remedy, and a cost estimate, plus a rating prediction and a marked-up script.
What do the report terms mean? (glossary)
Severities — BLOCKER: stops production, distribution, or insurance until resolved. HIGH: significant exposure; resolve before principal photography. MEDIUM: real but routine pre-production work. LOW: minor; fix opportunistically. FYI: no action required — on the record so the log shows it was considered.
Remedy verbs — CUT: remove the element or scene from the script (the zero-cost fix when the beat isn't load-bearing). REPLACE: swap for a cleared or fictional alternative. OBTAIN LICENSE: negotiate permission from the rights holder. OBTAIN RELEASE: signed release for a location, person, or artwork. RESHOOT: solve it in production — reframe, day-for-night, VFX. ADD DISCLAIMER: the standard on-screen disclaimer, an adjunct rather than a cure. ADD SPECIALIST: hire the named professional (stunt coordinator, armorer, animal handler). NO ACTION: considered and clear.
Music — sync license: permission to pair the composition with picture. master use: permission for a specific recording; a released track needs both, from different owners. PD: public domain — in the US, works published before 1930.
Verification — SUPPORTED: the citation backs the claim. PARTIAL: partly backed; read the caveat. REJECTED: the independent verifier found the source does not support the claim — the finding is struck and never affects the score.
These definitions also appear as tooltips on the chips inside every report.
Is this legal advice?
No. SCRIPTRISK is a research tool. It surfaces risks with sources so you and your attorney can act on them faster — it does not replace clearance counsel, and its output should not be relied on as legal advice.
Where do the citations come from?
Live web research at analysis time. Every finding quotes its sources verbatim — never paraphrased — with links. A finding the system cannot support with a source is not shown. Rating comparables quote each film's official CARA rating rationale, from the MPA's own register.
What is the verification step?
Every finding is re-checked by an independent verifier that sees only the claim, the scenes involved, and the cited sources — not the reasoning that produced it. Findings whose sources don't hold up are rejected, and the report shows you each rejection with the verifier's reason. In a typical run, a handful of findings don't survive.
What is the Greenlight Score?
A 0–100 number computed deterministically from the severity of surviving findings — each finding multiplies the score down, and a BLOCKER (something that makes the script unshootable or unreleasable as written) costs by far the most. The same findings always produce the same score.
How is the rating predicted?
By evidence, not vibes: your script's content profile is compared against 4,733 official CARA rating rationales from the MPA's own register. The prediction shows your nearest comparables and their actual ratings — and, if you're targeting a lower rating, the specific beats to change.
How accurate are the cost figures?
They are industry rules of thumb, not quotes, and the report labels them as such everywhere. They exist to tell you whether a fix is a $500 problem or a $50,000 problem.
What formats can I upload? How long does it take?
Fountain (plain-text screenplay format) today; PDF support is planned. A short screenplay takes 10–15 minutes and a feature 20–35 — the desks genuinely investigate and verify every citation, and they never cut corners to be fast. You can watch them work live while it runs.
Is my screenplay safe? How is it stored?
Your screenplay travels over HTTPS and is encrypted at rest with an application-held key — on top of Google Cloud's own storage encryption — so storage access alone yields ciphertext. It is never used to train AI models. Reports and script text are kept (encrypted) so your report link keeps working; report URLs are unguessable. If you're signed in, you can permanently delete any report and its script text from My reports at any time; for anonymous uploads, contact us and we'll remove them.
How is this different from Prescene, The Black List, or Filmustage?
Those tools answer "is this script good?" or "how do we schedule the shoot?" ScriptRisk answers "what will this script cost to shoot and release" — and it's the only tool where every finding carries a verbatim citation and an independent verifier rejects claims the source doesn't support. See the full comparison ›
Which territories are covered?
US, UK, China (theatrical), and UAE. That covers the major censorship regimes a typical release plan has to survive; more territories are on the roadmap.