Introduction to CID222

CID222 is a privacy-first AI gateway that protects sensitive data flowing between your applications and Large Language Models (LLMs).

What is CID222?

CID222 acts as a secure intermediary between your applications and AI models. It automatically detects and masks sensitive information before it reaches the LLM, ensuring your data stays protected while still enabling powerful AI capabilities.

Unlike traditional security tools, CID222 is purpose-built for the unique challenges of AI systems—understanding context, detecting PII in natural language, and preventing prompt injection attacks.

Key Features

• Automatic PII/PHI Detection — Identifies and masks 21+ entity types — names, emails, phones, national IDs, credit cards, IBANs — plus a DLP layer for leaked secrets (API keys, passwords, private keys).
• Content Safety Pipeline — ONNX safety models flag 13 toxicity categories, hate speech, and prompt-injection / jailbreak attempts before they reach your AI models.
• Policy-Based Access Control — Define custom rules for how different types of sensitive data should be handled, with support for masking, blocking, or flagging content.
• Complete Audit Trails — Every request, detection, and action is logged for compliance reporting and security analysis.
• Multi-Tenant Architecture — Isolate different teams, departments, or customers with separate configurations and credentials.
• Image & Document Analysis — Extract and redact PII from images (OCR) and PDF/DOCX/TXT/CSV documents, inline in chat or via dedicated endpoints.
• Real-Time Streaming — Full SSE (Server-Sent Events) support for streaming responses with sub-150ms latency overhead.

How It Works

CID222 intercepts all requests between your application and AI providers. Here's the flow:

1. Your application sends a request to CID222
2. CID222 analyzes the content through its safety pipeline
3. Sensitive data is detected and masked according to your policies
4. The sanitized request is forwarded to the LLM provider
5. The response is analyzed for any leaked sensitive information
6. The clean response is returned to your application

# Send a chat request through CID222
curl -N -X POST https://api.cid222.ai/chat/completions \
  -H "Authorization: Bearer cid_key_your_api_key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o",
    "messages": [
      {"role": "user", "content": "My email is john@example.com"}
    ]
  }'

# CID222 masks the email before forwarding, and streams the
# answer back as Server-Sent Events

Supported AI Providers

CID222 works with all major LLM providers:

• OpenAI (gpt-4o, gpt-4o-mini, gpt-4-turbo, gpt-3.5-turbo)
• Anthropic (Claude Opus 4.6, Sonnet 4.6, Haiku 4.5)
• Google (Gemini 2.5 Pro, Gemini 2.5 Flash)
• Azure OpenAI Service

Next Steps

Ready to get started? Head to the Quickstart Guide to set up CID222 in under 5 minutes.