Self-Hosted · Open Deployment

Stop AI data leaks
before they leave
your perimeter.

BitDrip inspects supported AI traffic your team sends to ChatGPT, Claude, Gemini and other LLMs — where configured policies can detect and block PII, credentials, PHI, and proprietary data in supported paths. The current product is for organizations and their authorized workforce; it is not currently offered for personal or consumer use.

SHA-256: 895c1411756c1711f441824d6030a32a0b6c7939d74ca203643a216fa4f27217  · 3 MB
Designed to support compliance with GDPR HIPAA PCI DSS SOC 2 ISO 27001 CCPA
YOUR DATA STAYS YOURS.
< 5 min
Average deploy time
6
Compliance frameworks
15+
Detection categories
< 10ms
Avg policy check time
How It Works

Supported AI traffic. Evaluated before it leaves.

BitDrip's proxy and policy components run in your network. A lightweight proxy daemon inspects configured supported HTTPS traffic, the policy engine evaluates it against your rules, and the configured action is applied before forwarding. The separate commercial portal manages licensing and release delivery.

YOUR NETWORK PERIMETER WORKSTATIONS user device user device HTTPS BITDRIP Proxy Daemon Port 3001 · TLS 1.3 · CA lifecycle Policy Engine 29 rules · 6 frameworks ✕ BLOCKED + LOGGED PII · PHI · Credentials detected ✓ allowed ↓ audit log · dashboard · SIEM export AI SERVICES ChatGPT api.openai.com · chat.openai.com Claude api.anthropic.com · claude.ai Gemini, Copilot + 12 more Supported AI providers monitored Perplexity · Mistral · Cohere · Grok · and more
Violation Detected — Blocked
When a supported configured detector returns a block action, the proxy rejects the request before forwarding. Validated audit configurations can record the rule, category, user identity, and a content hash without retaining the matched content in that audit record.
Clean Request — Forwarded
When policy allows a supported request, it continues to the AI service. Validated audit configurations can record bounded metadata; confirm deployed fields and retention before relying on this description.
Responses Scanned Too
AI responses pass back through the proxy. BitDrip checks for hallucinated PII, data exfiltration artifacts in responses, and jailbreak success signals on the return path.
Why BitDrip

Built for enterprise security teams

Nine capabilities that make BitDrip the right choice for organisations that take data privacy seriously.

Real-time Protection

Supported traffic to ChatGPT, Claude, Gemini and other configured AI services is checked before it leaves the device. Sensitive data — names, medical records, card numbers, passwords, and proprietary content — can be detected and blocked in real time. Tested browsers and applications that use the system proxy are protected with no browser add-on required.

AI Response Scanning

On supported configured response paths, BitDrip can evaluate documented PII patterns, jailbreak-success signals, and data-exfiltration artifacts. Coverage and detector outcomes require deployment-specific validation; no feature-parity claim is made about other products.

Customer-Controlled Content Path

The proxy, policy engine, dashboard, and their content-processing path run in your infrastructure. Supported configured requests go from the device through those customer-controlled components to the AI provider. The separate commercial portal is not in the documented prompt-content path.

Compliance & Audit

Built-in profiles and technical-control mappings can support an organization's GDPR, HIPAA, PCI DSS, SOC 2, ISO 27001, and CCPA work. Validated audit records can use hash-chain and signature evidence where the deployed release provenance confirms those features are enabled; the documented verifier can detect changes when verification runs.

SIEM Integration

Export audit events to Splunk (HEC/CEF), Elastic/OpenSearch, and Azure Sentinel in real time. Configurable per organisation — violations land in your existing security toolchain automatically.

MCP Policy Intercept Preview

Evaluate early MCP tool-call traffic through the policy layer for selected agent workflows. Full agentic AI discovery and monitoring remains on the product roadmap while the preview matures.

Throughput & Usage Analytics

Track AI usage volume by organisation and time period. Per-GB scanning metrics and time-series charts give security teams visibility into AI exposure across the entire organisation.

Fleet Visibility

Enroll workstations with a one-time token. The admin dashboard shows enrolled devices reported through the supported heartbeat path — including hostname, OS, proxy version, and online/offline status. Revoke a device in one click if it's lost or decommissioned.

System Tray App

A native app for macOS, Windows and Linux lives in your system tray and manages the BitDrip proxy daemon without any terminal required. Supports system proxy mode for centrally managed devices and PAC-based routing for BYOD and unmanaged devices. Shows live connection counts and blocked-request totals in the tooltip, with a built-in preferences window for proxy mode, policy engine URL, and CA certificate status.

Coverage & Compatibility

Know what BitDrip protects

Coverage depends on how each application connects. Validate every managed application before rollout.

Supported now
Tested browsers and applications that use your system proxy or the configured BitDrip PAC URL.
Requires configuration
CLI and SDK tools must be configured to use the BitDrip HTTPS proxy and local CA, then tested by your administrator.
Not guaranteed
Applications that bypass proxy settings, use certificate pinning, connect by direct IP, or use unsupported transports.

Review compatibility and validation guidance →

Security-first architecture

Covered routes and controls are implemented and tested to documented scope. Defense-in-depth is continuously assessed as the product evolves.

A01
Broken Access Control
Covered organization-scoped routes enforce and test tenant authorization. Authorization boundaries remain part of ongoing security review.
A02
Cryptographic Failures
AES-256 at rest, TLS 1.3 in transit, ed25519 license signing.
A03
Injection
Covered data paths use parameterized queries and schema validation. Covered output paths apply tested contextual encoding and sanitization, with defense-in-depth reviewed continuously.
A04
Insecure Design
Customer-controlled content processing and data-minimizing audit designs. Validate deployed fields, logs, diagnostics, and retention rather than assuming universal non-storage.
A05
Security Misconfiguration
Helmet.js (HSTS, CSP, X-Frame-Options). CORS allowlisting. No default credentials.
A06
Vulnerable Components
Automated dependency scanning. Dependabot-enabled weekly updates. Reproducible builds with npm ci.
A07
Auth Failures
Magic-link auth — no passwords, no credential stuffing. Short-lived signed JWTs. OIDC/SSO for enterprise deployments.
A08
Integrity Failures
SHA-256 installer checksums. ed25519-signed license JWTs. Pinned lock files.
A09
Logging Failures
Structured JSON audit logging. Sensitive fields sanitized at source. Tamper-evident event trail.
A10
SSRF
No external URL fetching from user content. Outbound connections allowlisted.

Zero Trust Data Flow

Validated audit configurations can use hashes and bounded metadata instead of full matched content. Confirm the deployed schema, logs, diagnostics, and retention before relying on that behavior.

Cryptographic Verification

Authenticated release evidence includes SHA-256 checksums for promoted installer bundles. License JWT verification uses ed25519; verify the deployed release and issuer evidence before use.

Self-Hosted by Design

Policy decisions for supported configured traffic run in your environment. The commercial portal has a separate, bounded account, licensing, deployment, billing, and release-delivery data path described in the privacy notice.

Designed to support compliance with GDPR HIPAA PCI DSS SOC 2 ISO 27001 CCPA

Documented, tested capabilities can support selected technical-control objectives on validated paths; this is not certification or complete control coverage. Read the documentation →

Compliance Coverage

Designed to support your compliance requirements

BitDrip provides configurable detectors, policy actions, and evidence that can support selected framework objectives on validated paths. It does not determine legal compliance or guarantee detection or blocking.

GDPR
PII & special-category data — names, emails, addresses, national IDs
HIPAA
PHI & patient data — medical records, conditions, insurance IDs
PCI DSS
Card numbers (Luhn-validated), CVV codes, PINs
SOC 2
API keys, passwords, secrets & credentials
ISO 27001
Credentials, proprietary data & internal document markers
CCPA
California resident PII & consumer personal data

Designed to support compliance — not a substitute for legal review.

Getting Started

From zero to protected in minutes

Register once, deploy anywhere. No SaaS. Your AI data stays in your infrastructure, always.

1
Register
Create your account and choose a plan. Receive your signed license file by email.
2
Download
Download the installer bundle. Verify the SHA-256 checksum. Extract and review.
3
Deploy
Run ./install.sh — it starts all services automatically. Then run bitdrip cert generate && bitdrip cert install to install the CA certificate, and bitdrip proxy start to begin intercepting traffic (or use the system tray app).
4
Protect
Configure policies in the dashboard. Enroll additional workstations from Settings → Workstations. Test browsers and configure each CLI or SDK tool that must use the BitDrip proxy.

Full installation guide, configuration reference, and troubleshooting are in the documentation.

Read the Docs →
See It In Action

Visibility into supported AI interactions

The dashboard presents policy events, technical-control mappings, and enrolled-workstation state from the customer-controlled deployment. Validate deployed audit fields, logs, diagnostics, and retention; the dashboard does not establish legal compliance.

BitDrip Dashboard · Recent Violations
23 blocked today
2,841 allowed today
12 workstations
PII
Full Name detected → ChatGPT
"Please analyze patient John ████████'s lab results..."
hr@company.com · 2 minutes ago · BLOCKED
PHI
Medical Record ID detected → GPT-4
"Patient MRN ████████ was diagnosed with..."
clinic@company.com · 8 minutes ago · BLOCKED
CRED
API Secret Key detected → Claude
"sk-████████████████ is our prod key for..."
dev@company.com · 14 minutes ago · BLOCKED
PII
Email Address detected → Gemini
"Contact ████████@client.com to close the deal..."
sales@company.com · 31 minutes ago · BLOCKED
BitDrip Dashboard · Compliance Overview
94%
GDPR
87%
HIPAA
100%
PCI DSS
96%
SOC 2
91%
ISO 27001
98%
CCPA
LAPTOP-MBP-01 macOS · Client 1.3.3
DEV-UBUNTU-03 Linux · Client 1.3.3
WIN-HR-WORKST-04 Windows · Client 1.3.3
FINANCE-WIN-02 offline
Pricing

Simple, transparent pricing

Start free. No credit card required for the Community tier.

For licensing, one user means one enrolled workstation or device seat.

Community
Free
evaluation & lab use
  • Up to 3 users
  • Supported detection rules
  • HTTPS proxy daemon + system tray app
  • Audit logging
  • Community support
Download Free ↓
Team
$99/mo
month-to-month  ·  $83/mo billed annually ($999/yr)  SAVE 16%
  • Up to 10 users
  • 1 deployment
  • Supported detection rules
  • HTTPS proxy daemon + system tray app
  • Audit logging
  • Email support
Get Started →
Starter
$299/mo
month-to-month  ·  $250/mo billed annually ($3,000/yr)  SAVE 16%
  • Up to 50 users
  • 1 deployment
  • Supported detection rules
  • HTTPS proxy daemon + system tray app
  • Compliance reports (CSV + printable report)
  • Email support
Get Started →
Enterprise
Custom
pricing
  • 500+ users
  • High-volume deployments
  • Kubernetes / Helm deployment
  • HA guidance; air-gap by engagement
  • CA lifecycle management & audit log
  • Custom SLA & onboarding
  • Dedicated CSM
  • MDM / GPO cert deployment
  • Volume licensing
Contact Sales →
Ready to stop the leaks?
Register free, download the installer and have BitDrip running in your environment today. Community tier covers up to 3 users for evaluation. Team plan starts at $99/mo for up to 10 users.
Linux x86_64 Linux arm64 macOS Windows x64 Docker Podman
Platform version: v2.2.58 ·  Released: 2026-07-23 ·  SHA-256: dfdf951fb0a4000c7ef92c52b2d728b2710ac2c9ad231719193f426dfba5e40a
More from Anchor Cyber Security