Your script's privacy

Your script is yours. Period.

We know exactly what feels risky about uploading a screenplay to a platform that writes coverage automatically: will it be used to train a model? Will pieces of it end up in someone else's script? Who else can read it? Here is the full answer, with no fine print.

What happens when you request coverage

1
Your script is read in full, once.
The reading runs on a screenplay-specialized system built on Anthropic's Claude models, with professional criteria: structure, character, dialogue, tone.
2
The coverage is generated, and that's the end of it.
The system has no memory between reads: your script isn't "learned", doesn't feed any training, and doesn't influence anyone else's coverage.
3
The result lives only in your workspace.
The coverage and the script file stay encrypted in your private space, visible only to your team — and to whoever you choose to invite.

What never happens to your script

It is never used to train models. That's prohibited by the model provider's commercial terms — a contract, not a promise.
It is never used to write or improve other scripts. Every read is isolated.
It is never shared with other customers, third parties, or anyone outside your workspace.
The SlateCore team doesn't read it. Administrative access is restricted to support you request, and it's audited.

So why is the coverage this good?

Because the language model is just the engine. The professional reader is what we built — and that's where the specialization lives:

A professional evaluation rubric. The same criteria a development department uses to evaluate material — concept, structure, character, dialogue, tone — with per-section scoring, not a general opinion.
Verdicts calibrated against a real corpus. A "Recommend" doesn't come from the model's intuition: it's calibrated against the distribution of hundreds of real coverage reports, so it means the same thing today as it will in six months.
Coverage format and craft. It comes out in the industry's standard format, anchored to specific pages and scenes of your script — not generic chatbot paragraphs.
Continuous quality control. We monitor the system's score and verdict distributions to detect and correct any drift.

None of that requires training on customer scripts. The specialization lives in the reading system, not in your data.

What about pasting it into a free chatbot?

Consumer chatbot (default settings)

Your conversation may be used to train the model.
The script sits in a personal chat history, outside your company's control.
No watermarks, no access logs, no NDA.

SlateCore

Processing under commercial terms that prohibit training on your content.
The script lives encrypted in your company's workspace, with roles and permissions.
If you share it, it's through data rooms: individual invites, per-person watermarks, signed NDAs, and a log of who saw what.

How we guarantee it

Contract, not trust. Coverage is processed with Anthropic's models under their commercial terms for API customers, which prohibit using customer content to train models. Their data processing agreement is public: Anthropic's Data Processing Addendum.
Transient processing. The model provider doesn't keep your script for any other purpose: only temporary operational records, deleted under their retention policy.
Private, encrypted storage. Your script is stored in private buckets (no public URL), encrypted in transit (TLS 1.3) and at rest (AES-256), with per-company isolation at the database level.
You control the lifecycle. Versions under your control, access revocable at any time, and deletion whenever you decide.

The full technical detail — encryption, backups, subprocessors, access policy — lives in Security and the Privacy Policy.

Straight questions, straight answers

Does SlateCore train an AI on my script?

No. Neither SlateCore nor the model provider trains anything on your content. Coverage uses general-purpose models under a no-training contract, with a screenplay-reading system built by SlateCore on top.

Can my script show up in someone else's coverage or screenplay?

No. Every read is isolated: the system has no memory between requests and your script is not part of any model or any other workspace.

Who can read my script?

Your team, according to the roles and permissions you configure — and the external guests you invite to a data room, each with their own watermark, signed NDA, and access log. No one else.

What happens if I delete my script?

It's removed from your workspace. The platform's encrypted backups rotate and expire under the retention policy published in the Privacy Policy.

Professional coverage without giving your material away

That's the whole deal: your script goes in, the coverage comes out, and the script stays yours.

Try SlateCore

Need more detail for your company's evaluation? Write to security@slatecore.app