It reads the whole thing before it says a word about it.

No. Scripitt does not generate creative material of any kind. It reads what is submitted and produces an assessment of it. The writing stays the writer's.

Minutes rather than days for a standard feature or pilot. Longer manuscripts take proportionally longer because the whole work is genuinely read before anything is judged. We position on speed, not on a contractual turnaround.

It is stored against your organization's workspace under row-level access control, visible to your team and whoever you share it with. Scripitt does not use submitted scripts, manuscripts or other customer material to train Scripitt models. Scripitt-managed analysis runs on a single approved AI provider whose commercial terms do not permit training on customer inputs, and no second provider is used unless your organization asks for it. Projects can be deleted by their owner, and share links can be revoked at any time or set to expire. Enterprise organizations may also run Scripitt on their own approved AI-provider account, subject to the data-handling terms of that provider.

No. A general-purpose model handed a script produces an opinion. Scripitt produces a governed evaluation: a complete comprehension pass, parallel specialist analysis reconciled rather than averaged, a validation ladder that rejects and repairs weak output, weighted scoring, provenance, revision tracking and a recorded company decision above the machine's recommendation. The model does the reading at defined points inside that process. The process is the product, and it is what the pending patents cover.

One tested configuration, the same for every workspace: a single approved provider, with one model for judgment and one for support work. Scripitt does not publish model version names, because the evaluation framework is designed to outlive any particular model. An enterprise can run on its own approved provider account instead.

Anyone holding the link, which is why links are created only by authorized users, can be revoked at any time, and can be set to expire after 7, 30 or 90 days. A shared report carries an organization's branding only when the person who shared it belongs to that organization's workspace; it cannot be dressed in someone else's identity.

The creator keeps ownership of the submitted material; Scripitt acquires no copyright or authorship in it, and does not acquire authorship of the evaluations either. Reports are advisory analysis produced for the organization that commissioned them. Nothing the platform produces determines rights, credit or compensation.

Yes. Pitch documents, series treatments and film treatments each have their own 100-point rubric, so they are judged as what they are: never marked down for missing sluglines, scene dialogue, act breaks or screenplay formatting, and the report says plainly that those elements were not evaluated. Gaps that do matter, such as an unclear series engine, still count. Scores are specific to the format: a strong pitch is measured against pitches, not against a finished pilot. The document type is detected on upload with the reason shown, and if the platform is unsure it asks you to choose rather than guess.

No. It is an internal chain-of-custody record — a fingerprint and a timestamped chronology. It is not a registration, not proof of authorship and not a credit determination.

Yes. Point a lens at material your team responded to, or describe the taste in your own words, and it calibrates against that sensibility, so evaluations reflect how your organization reads rather than a generic industry average.

PDF, DOCX, legacy DOC, TXT, RTF, Markdown, Final Draft (.fdx) and Fountain. Scanned PDFs are handled with vision OCR.

Only in Topical Coverage. That mode checks time-sensitive material against current web sources and cites what it found so a person can verify it. The other ten modes evaluate the submitted material on its own, without live-web research. In every mode the recommendation is the platform's and the decision is yours.

No. The recommendation is the platform's; the decision is yours, and the two are recorded separately. When your team logs a company decision, the brief and the library register show who made it and when, alongside — never instead of — what the evaluation said.