Paste or load text
Bring in the material you plan to use with AI.
MUTANT Data Safety Layer
Stop sensitive databefore it reaches AI.
The Data Safety Layer checks text locally in your browser, detects and redacts 11 types of sensitive data, and supports privacy-preserving audit evidence—before you share anything with an AI tool.
- Local-first checking
- 11 sensitive-data types
- No account required for the web app
Text entering AI workflowOn device
Hi, please summarise the case for Sarah Morgan. Her email is sarah@example.com, phone 07700 900123, and the reference is AB12 34CD 56EF.
4items detected
11types checked
0text uploads
The safety flow
Four steps. Your text stays on the page.
A clear checkpoint between sensitive working material and the AI tools your team uses.
Detect sensitive data
Check for 11 categories, including contact, identity and financial details.
Review and redact
Inspect the findings and replace sensitive values with clear placeholders.
Export a safe record
Copy safer text and, when needed, create privacy-preserving audit evidence.
Scanner demonstration
See what the Data Safety Layer catches.
Hi, please summarise the case for [PERSON]. Their email is [EMAIL], phone [PHONE], and the reference is [IDENTIFIER].
Illustrative example. Detection results should always be reviewed before sharing text onward.
One practical AI toolkit
Create. Protect. Verify.
Move from better inputs, through safer handling, to stronger checks on what you see and share. Every route below opens a real tool or product.
Create
Sharper prompts and experimental creative provenance.
Chrome extension
Prompt Power Pro
Save, organise and reuse your best prompts across AI workflows instead of rebuilding them from scratch.
Explore Prompt Power Pro → Open-source R&DComfyUI-Mutantwork
Experimental ComfyUI nodes exploring creator identity and provenance metadata in image workflows.
View the GitHub project →Protect
Local-first checks before sensitive text reaches AI.
Web app
Data Safety Layer
Detect and redact 11 types of sensitive data locally, with support for privacy-preserving audit evidence.
Open the Data Safety Layer → Chrome extensionData Safety Layer for Chrome
Bring local sensitive-text checks into your browser workflow. See the product page for current features and access details.
Explore the Chrome extension →Verify
Practical tools for image and text provenance signals.
Web app
MUTANT Verify
Inspect images using available authenticity and AI-detection signals, with results presented for human review.
Scan an image → Local-first web appMUTANT Inspect
Read image provenance signals, Content Credentials and privacy-sensitive metadata locally in your browser.
Inspect an image → Chrome Web StoreVeriText
Estimate whether text may be AI-generated and review extracted factual claims with confidence scoring before you share it.
Get VeriText →Why MUTANT
AI tools built differently.
Small, understandable products designed around control—not another black box between you and your work.
Keep sensitive work close
Where a task can run in your browser, it should. The safety web app is built around that principle.
Show the checkpoint
Useful output should help you review and decide, not hide judgement behind a score.
Built without platform lock-in
Focused tools that can fit alongside the AI services you already choose.
Made for real workflows
Direct routes, clear interfaces and tools meant to be used—not just demonstrated.
Built independently
Built by Rick.
“MUTANT is where I build practical, opinionated tools for working with AI more safely, creatively and transparently.”
Rick Brereton · Product builder
Open source & R&D
Experiments and special projects.
The public workbench: open-source creative tooling and focused products for overlooked operational problems.
ComfyUI-Mutantwork
Experimental custom nodes for ComfyUI, exploring creator identity and provenance metadata in image generation workflows.
View on GitHub →NI Trade Guard
A specialist project for navigating GB-to-NI shipping paperwork and Windsor Framework checks. Open the current hosted app for its latest scope.
Open NI Trade Guard →Make safety the step before send
Check your text before sharing it with AI.
Start in the browser with a local-first check. Review what is detected, redact what matters, then move forward with more control.