Draftbox

Privacy is not a setting. It is the shape of the space.

Create visibility zones inside your Workspaces — even inside a single document. Control who sees what, down to the sentence.

Zones of visibility.

In most platforms, sharing is all-or-nothing. You share a document, and everything inside is visible. Draftbox lets you create visibility zones within Workspace Sections — areas that only specific people can see. A team member might see the analysis but not the raw data. A collaborator might see the conclusion but not the draft that led to it. The same document, different worlds.

Full visibility

Sees everything in the section

Restricted zone

Sees only what is explicitly shared

Confidential block

Invisible to everyone except the author

The mask tool.

Select any part of your text. Mark it as confidential. It becomes invisible to collaborators, to AI processing, to external APIs. The words stay in your document — but they travel nowhere you do not send them.

From you, visible

You see the masked text in your editor. It is your words. You wrote them. You control them.

From AI, hidden

Masked sections are stripped before any AI call. The model never sees them. They do not enter the RAG pipeline. They cannot leak through responses.

From others, gone

Collaborators see a placeholder where the text was. They know something is there. They cannot read it. Permission is yours to grant, not theirs to take.

Names, emails, phone numbers

Credit cards, bank accounts, IBANs

National IDs, passport numbers, tax IDs

API keys, tokens, passwords

IP addresses, URLs with credentials

Dates of birth, home addresses

Full sovereignty, if you need it.

In the air-gapped deployment mode, none of it leaves your infrastructure — every AI call, embedding and rerank runs on your own models. On the roadmap: a mask tool to mark any part of a document as confidential, delivered to the AI as [CONFIDENTIAL].

Your data is not training material. Your thoughts are not a product. What you think here stays here.

Your data. Your rights. Your control.

GDPR isn't a checkbox. It's a commitment. You can access, correct, export, restrict, or delete your data — at any time, no questions asked.

Right to access & port

Export everything you've ever created in a structured, machine-readable format. Your data is portable — it goes where you go.

Right to be forgotten

Request complete erasure of your data. The system processes deletion requests within mandated timeframes — no retention beyond what the law requires.

Right to rectify & restrict

Correct inaccurate data or restrict processing at any time. You decide how your information is used — not the platform.

Data Subject Access Requests — handled, not ignored

Every data request is tracked from submission to completion, with automated workflows ensuring GDPR-mandated response timeframes are met. Request history is retained for audit compliance — so regulators can verify, and you can trust.

Every piece of data knows what it is.

Content is automatically classified by sensitivity — from public information to restricted material requiring special handling. This classification drives access controls, encryption policies, and retention rules. Nothing is left unclassified. Nothing is treated the same.

Public

Freely shareable, no restrictions

Internal

Organization-wide, not external

Confidential

Need-to-know, access controlled

Restricted

Highest sensitivity, logged access

Retention with purpose

Data is retained only as long as legally required and no longer. Security audit logs are kept for 7 years to meet compliance obligations. Data subject records are purged automatically once retention periods expire — the system cleans up after itself.

Automated enforcement

Retention policies aren't just documented — they're enforced. Daily automated processes classify unclassified data and purge expired records. And the purge itself is logged, so auditors can verify compliance.

Control who sees what. Down to the sentence.