About WorkLens
Your data should never leave your hands
WorkLens is a browser-based de-identification platform that doesn't just fake rows — it understands how your data is connected, calculated, timed, and validated. One engine, three modes: generate synthetic data, redact PII, or apply deterministic morphing. No server. No upload. No cloud. Files are processed and downloaded without a single byte leaving your machine.
The Problem
You need to share a spreadsheet — but it contains names, tax file numbers, and phone numbers.
You could manually search and redact. For 10,000 rows, that takes hours and you will miss things.
You could upload it to a cloud tool. Now a third party has your sensitive data.
You could write a script. Now you need to maintain regex patterns for every country and ID format.
How WorkLens Solves It
Same engine, three modes — pick one based on what your file needs. Generate fake data from a recipe (no real file required), or drop a real file to redact or morph.
Pick a mode
Synthesize, redact, or deterministic morph. Same engine, same privacy promise.
Drop a file or recipe
CSV, XLSX, PDF, JSON, or XML. Or start from a *.wl.json recipe — no real file needed.
15 layers process it
Regex, checksums, AI names, semantic headers — all in your browser.
Download + receipt
One click. PII replaced or new rows generated. Signed proof for every job.
What Runs Where
Every component of WorkLens runs inside your browser tab. Nothing is hosted, proxied, or relayed through a server.
Finds patterns like emails, phone numbers, and ID formats. Validates numbers mathematically.
On-device AI that recognises person names, organisations, and addresses. Downloaded once, cached for instant future use.
Understands what your columns contain by analysing headers against known data patterns.
Reads CSV, XLSX, JSON, and XML natively. Your data never leaves the browser.
Heavy work runs in the background so the interface stays responsive during scans.
15-Layer Detection Pipeline
Each layer is a separate module. They run independently and results are fused by confidence — no single layer can override another without evidence.
Greyed out layers are planned for future releases.
By the Numbers
Built For
Compliance Teams
Audit spreadsheets for PII before sharing with third parties. Cryptographic audit trail for every scan.
Financial Services
Banks, insurers, and regulated entities handling customer data. Mathematical validation for tax IDs and government numbers.
Legal Professionals
Redact client names, addresses, and identification numbers from discovery documents and case files.
Healthcare Providers
Detect patient identifiers, health insurance numbers, and personal details in clinical data exports.
Data Engineers
Scan datasets before loading into analytics pipelines. Catch PII that slipped through upstream processes.
Anyone With a Spreadsheet
If you have a CSV with names and you need to remove them before sharing — this is the fastest way.
What's Coming
- CSV, XLSX, JSON, XML scanning
- AI-powered name detection
- 20+ PII types with mathematical validation
- One-click redacted download
- Multi-select bulk redaction
- Headerless file support (per sheet)
- Digitally signed audit receipt
- Australia (deep), universal patterns (global)
- PDF scanning with bounding-box redaction
- Deterministic morphing — realistic or token-based
- APRA CPS 234 compliance policy packs
- Human review queue for low-confidence items
- Desktop app via Tauri (macOS + Windows)
- TEE attestation for regulated industries
- Format-preserving encryption (FF1)
- Python SDK via PyPI
- US, UK, EU, NZ region support
See it for yourself
No signup. No credit card. Drop a file and watch it work.