Trust & data boundaries

Trust starts with where the work happens.

Redactum is a local-first application for privacy-sensitive document workflows. Your original PDF is read, detected, and redacted on your machine—before you decide whether a sanitised artifact belongs in an AI workflow.

Local-first architecture

Redaction is a local operation, not a cloud upload.

That separation is the point: prepare a safe copy locally, review it, then choose what happens next.

YOUR LOCAL ENVIRONMENT

Everything in this journey happens on your machine.Original document → local redaction → reviewable output

Redactum local workspace showing a demo service agreement before redaction. — 01 · BEFORE REDACTIONOriginal stays local.

Redactum local workspace showing the same demo service agreement with personal data redacted. — 02 · REDACTED OUTPUTReview the output before it moves on.

OPTIONAL, CUSTOMER-CHOSEN HANDOFFRedacted PDF or sanitised text

Your AI, search, or extraction service

Trust centre

Clear controls. Clear choices.

We describe the product as it works today, so you can judge whether it fits your organisation and workflow.

Local processing

Redactum is a local application, not a document-processing cloud SaaS. Document reading, OCR where needed, sensitive-data detection, and PDF redaction run in your environment.

No original-PDF upload for the redaction path
Digital and scanned PDF support
Packaged runtime for local processing

Output you can inspect

A redaction run creates the artefacts needed to review what happened before the document is used elsewhere.

Redacted PDF and sanitised text
Detection details and content identities
Verification report when that optional step is used

Controlled handoffs

Redactum does not automatically send a document to an AI provider. You decide whether a sanitised artifact is used in a downstream AI, search, or extraction workflow.

Your chosen service, not a built-in document cloud
A selected sanitised artifact leaves only when you choose
The original document remains inside the local flow

Workflow-specific implementation

We start with the workflow you actually need to make safer. Early customers help shape integrations, controls, and the product roadmap around real document work.

Detailed workflow review before a trial
Tailoring discussed around your environment
Direct feedback loop with the people building Redactum

Optional contract verification

Use the LLM provider you choose.

Contract verification is a separate, explicitly configured step. Redactum sends sanitised contract text and an approved reference template to the OpenAI-compatible LLM endpoint your organisation selects, then saves a local verification report.

This gives your team a choice: keep redaction entirely local, or enable verification with a provider that fits your own policies and contracts.

Sanitised contract text
+ approved templateYour configured LLM providerLocal verification report

Redactum verification result showing two contract changes that require review. — EXAMPLE RESULTSurface changes that require review.

Built with customers

Your workflow is part of the product conversation.

We invest time upfront to understand the documents, tools, review steps, and data boundaries your team works with. The goal is not a generic AI pitch—it is a workflow that is genuinely useful and safe to adopt.

01
Map the real workflowLook at where documents come from, where they need to go, and what has to stay protected.
02
Shape the right rolloutDiscuss setup, handoffs, and any tailoring needed for your environment.
03
Improve it togetherEarly customer feedback directly informs Redactum’s product priorities.

Talk through your workflow