AI-Safe Content Sharing
Use AI tools safely with pre-redacted content
Outcome
Prepare reviewed redacted context before another AI tool sees it.
Your progress
Mark each step as you complete it. This is saved only in this browser.
0 / 3 steps complete
Before you start
Visual frame
Review asset
AI-safe content sharing storyboard
Synthetic assistant frame only. It keeps raw originals out of prompts and avoids training, storage, or output-guarantee claims.
Synthetic frame
Assistant context
Use the reviewed redacted version.
Instructional before-AI frame
A simplified handoff frame for preparing redacted context before using another AI tool.
Guided steps
Follow the task, then check the result.
Prepare only the context the AI tool needs
Trim unrelated material first. The less unnecessary private context you include, the easier it is to review the redacted result.
Redact and review before copying
Run detection, inspect placeholders, and confirm the redacted context still preserves meaning without exposing raw private values.
Copy the reviewed redacted context
Copy only the reviewed redacted text into the external AI tool. Do not copy the redaction map or raw original text as part of the prompt.
Branch questions
Before opening another AI tool, do you see reviewed redacted context?
Do you still see raw private values in the text you plan to send?
Completion check
Support boundary
Support can use by default
- Workflow step, selected symptom, and external handoff category.
- Counts by detection type, if available.
- Redacted sample only after explicit consent.
- Intent ID
- Article slug
- App route
- Browser and viewport
Requires your consent
- User-written description
- Email address
- Explicit attachment
- Redacted sample
- Support bundle previewed to the user
Not collected by default
- Raw original document text
- Full local file paths
- Pasted private content
- Unmasked screenshots
- Replay capture
Safe support summary
Copy a scrubbed handoff.
Redactorr KB support summary Article: /knowledge-base/ai-assistant Selected issue: You cannot produce reviewed redacted context from a supported input. Playbook progress: 0/3 steps marked complete Safe context to include: - Workflow step, selected symptom, and external handoff category. - Counts by detection type, if available. - Redacted sample only after explicit consent. - Intent ID - Article slug - App route - Browser and viewport Requires explicit consent: - User-written description - Email address - Explicit attachment - Redacted sample - Support bundle previewed to the user Do not include by default: - Raw original document text - Full local file paths - Pasted private content - Unmasked screenshots - Replay capture User note: - Describe the step and symptom without pasting raw document text, secrets, files, or restoration material.
Article details
Feed AI Without Feeding It Your Secrets
You want to use ChatGPT to summarise a contract. Or Claude to draft an email. Or Copilot to generate code. But the document contains client names, account numbers, and confidential terms.
AI-Safe Sharing solves this.
Redact first, then paste into AI. Your secrets stay private. The AI gets enough context to help.
How It Works
Redact your content: Run detection on your document. Sensitive data becomes redacted items like [EMAIL_1], [COMPANY_1], [ACCOUNT_1].
Paste into AI: Copy the redacted version. Paste it into ChatGPT, Claude, or any AI tool.
AI processes redacted items: The AI sees redacted items instead of real data. It can still understand structure and context.
Restore locally: When the AI returns results, use your redaction map to restore real values - but only on your machine.
What Gets Protected
Personal identifiers: Names, emails, phone numbers → [NAME_1], [EMAIL_1], [PHONE_1]
Financial data: Account numbers, credit cards → [ACCOUNT_1], [CARD_1]
Proprietary info: Internal IDs, customer codes → [ID_1], [CUSTOMER_1]
Confidential terms: Contract values, pricing → [AMOUNT_1], [PRICE_1]
Real-World Use Cases
Legal contract review: "Summarise this NDA but don't see the parties' names."
Customer support drafting: "Write an email response to this issue without seeing customer details."
Code generation: "Generate SQL queries without seeing real database credentials."
Financial analysis: "Analyse this P&L without seeing actual revenue figures."
Medical documentation: "Draft a referral letter without seeing patient names."
External AI Provider Integration
Bring Your Own Key: Use your own OpenRouter API key. Redaction happens before the request leaves your browser.
Supported models: GPT-4, Claude 3.5 Sonnet, Gemini Pro, Llama 3, Mixtral, and 100+ more.
Streaming responses: Get real-time results as the AI generates.
Restoration: Restore redacted values after AI processes - all client-side.
Privacy Boundary
Pre-redaction: Redact before sending reviewed content to AI.
Reviewed redacted context: AI workflows should use reviewed redacted context rather than raw original values.
Redaction maps stay local: Redaction maps are browser-side review artefacts and are not part of the reviewed redacted AI context.
Audit trail: See exactly what was sent to which AI model.
Example Workflow
Paste contract → "Party A (Acme Corp) agrees to pay Party B ($50,000)..."
Redact → "Party A ([COMPANY_1]) agrees to pay Party B ([AMOUNT_1])..."
Send to AI → "Summarise key terms: Party A ([COMPANY_1]) agrees..."
AI responds → "Summary: [COMPANY_1] will pay [AMOUNT_1] to..."
Restore → "Summary: Acme Corp will pay $50,000 to..."
Supported AI Providers
Via configured AI provider:
- OpenAI (GPT-4, GPT-3.5)
- Anthropic (Claude 3.5 Sonnet, Opus)
- Google (Gemini Pro, Ultra)
- Meta (Llama 3)
- Mistral (Mixtral, Mistral Large)
- Perplexity (Sonar)
Direct integration: Use any AI tool manually: ChatGPT, Claude.ai, Copilot, Bard. Just paste redacted text.
Still stuck?
Copy a safe article handoff.
Start support with the article, the issue, and safe context. Raw document text, files, restoration material, and unredacted screenshots stay out unless you explicitly choose otherwise.
- Article slug
- Selected issue
- Screen name
- Action name
Redactorr support case Source: Article playbook: ai-assistant Route: /knowledge-base/ai-assistant Selected issue: You cannot produce reviewed redacted context from a supported input. Safe context: - Workflow step, selected symptom, and external handoff category. - Counts by detection type, if available. - Redacted sample only after explicit consent. - Article slug - Selected issue - Screen name or article section - Action name - Browser and viewport Only include with explicit consent: - User-written description - Email address - Explicit attachment - Redacted sample - Support bundle previewed to the user Do not include by default: - Raw original document text - Full local file paths - Pasted private content - Unmasked screenshots - Replay capture - Hidden diagnostic uploads User note: - Describe what you tried and what happened. Do not paste raw document text, secrets, files, restoration material, or unredacted screenshots.
Support case builder
One case format, wherever you start.
This is the same support case shape used by diagnostics and article handoffs.
- Workflow step, selected symptom, and external handoff category.
- Counts by detection type, if available.
- Redacted sample only after explicit consent.
- Article slug
- Selected issue
- Screen name or article section
Private material belongs behind an explicit consent step, not in the initial case.
6 default exclusions