Step 1 — Configure
Your rules. Your data.
Blocklist
Add database passwords, API keys, internal URLs. Each entry is AES-256 encrypted. Exact string matching — if Bleep sees it, it catches it.
Patterns
Policies
Step 2 — In action
Three environments. Three responses.
A developer pastes a database connection string into ChatGPT. Bleep catches the credential and blocks the request.
An AI agent processes a terminal screenshot. Bleep extracts text via OCR, detects secrets, and redacts the image before forwarding.
Code with internal URLs is sent to Copilot. Bleep detects infrastructure references and logs a warning. The request goes through.
Step 3 — Detection engine
Five layers, every request.
All local. Nothing leaves your machine. Every layer runs in parallel.
API Keys & Tokens
AWS, OpenAI, Stripe, GitHub, 40+ formats. Unlimited custom regex.
sk-proj-K8x•••Sensitive Values
Exact-match blocklist. Passwords, secrets, internal values. AES-256 encrypted.
password → ••••Personal Information
SSNs, emails, phone numbers, credit cards. Detected and redacted.
482-33-••••Screenshots & Images
OCR extracts text from screenshots of terminals, dashboards, config files.
.env → scannedDocuments & Files
PDF, DOCX, XLSX, PPTX. Content extracted and scanned before upload.
.docx → 2 hitsStep 4 — Deploy
Your infrastructure. Your rules.
Desktop app for individuals. On-premise server for teams. Nothing leaves your network.
Cloud
Bleep Cloud
Account, billing, license keys, downloads
Your device
Bleep Desktop App
Standalone, self-contained, no server required
DNS redirect
Only AI traffic intercepted — zero config
Detection engine
Secrets, PII, API keys, blocklist
File scanning
OCR images, PDFs, Office docs
Local SQLite DB
Violations, policies, config — 100% on-device
Single device · 100% local processing · No server required
See it yourself in five minutes.
No code changes. No cloud dependency. Install, run, done.