The browser-native LLM firewall
Obfusca scans every prompt and file upload before it reaches AI tools like ChatGPT, Claude, Gemini, Grok, DeepSeek, and more. Sensitive data is detected, redacted, or blocked automatically.
11+ built-in detection patterns
Pattern-based detection catches SSNs, credit cards, API keys, and more using regex with validation. Plus, 24 built-in semantic rules protect your personal, financial, medical, and digital identity data. Each pattern includes a confidence score and optional validator to eliminate false positives. Four policy actions are available: block, redact, warn, or allow.
| Pattern | Category | Severity |
|---|---|---|
| US Social Security Number | PII | critical |
| Credit Card Number | PII | critical |
| Email Address | PII | low |
| AWS Access Key ID | Secrets | critical |
| AWS Secret Access Key | Secrets | critical |
| API Key (sk- prefix) | Secrets | high |
| API Key (generic) | Secrets | high |
| Bearer Token | Secrets | high |
| Private Key (PEM: RSA, EC, DSA, OPENSSH, PGP) | Secrets | critical |
| GitHub Token | Secrets | high |
| Slack Token | Secrets | high |
Custom detection patterns
Define your own rules beyond the built-in patterns. Use regex for format-based matching or keyword lists for exact terms. Custom patterns support four actions: block, redact, warn, or allow.
8 Pre-built templates
Semantic detection
Go beyond pattern matching. Full semantic detection reads context and catches what regex never could. Write custom rules in plain English, apply industry-specific presets, and protect your entire team.
Individual plan
Personal Usage preset with 24 built-in semantic rules covering identity, financial, medical, contact, and digital data. Catches spoken-out SSNs, partial card info, and context-dependent PII that regex misses.
Team plan
Full accessAll industry presets (technology, finance, healthcare, legal) plus up to 20 custom rules in plain English. Enterprise plans get unlimited rules. Write rules like “flag settlement amounts” or “detect internal project codenames” and Obfusca understands intent, not just keywords.
What semantic detection catches
File upload scanning
When users upload files to AI tools, Obfusca extracts text content and scans it with the same detection engine. Supports 70+ file types up to 10 MB.
Documents
PDF, DOCX, PPTX, ODT, RTF
Spreadsheets
XLSX, XLS, CSV
Code
PY, JS, TS, JSX, TSX, Java, SQL, Go, Rust, C/C++, and 30+ more
Config
JSON, XML, YAML, ENV, INI, TOML, Dockerfile
Text
TXT, MD, LOG
Notebooks
IPYNB (Jupyter)
Test before you deploy
Test any pattern, built-in or custom, against sample text before rolling it out to your organization. See exactly what matches, where it matches, and at what confidence level. No risk, no guesswork.
Central dashboard for security teams
Manage policies, review detection events, and understand where risky prompts and files were stopped. All from one place.
Policy Management
Configure block, redact, warn, or allow per detection type. Apply protection profiles that set up detection rules and policies in one click.
Event History
Paginated log of every detection event with filters by source, decision, and severity. Bypass incidents are flagged with full details for incident review.
Team Management
Invite members, assign admin or member roles, and monitor extension status per user — see who has protection active, disabled, or not installed.
Audit Log
17 tracked admin action types. See who changed what and when.
Ready to protect your AI usage?
Scan prompts and file uploads across 11 AI platforms. Detect secrets and PII in real time, then mask, redact, or replace sensitive data before it leaves the browser.