About Secure LLM Proxy
Secure LLM Proxy is an OpenAI-compatible proxy for routing model requests with PII scrubbing and audit logging. It also exposes chat completion as an MCP tool.
Key features
- 01Scrubs email, phone, SSN, and card data
- 02Logs request hashes for audits
- 03Routes supported models through one endpoint
- 04Pools upstream keys for rate limits
- 05OpenAI-compatible endpoint
Use cases
- 01Remove identifying details before forwarding prompts
- 02Review hashed request records
- 03Route requests across supported model providers
- 04Distribute load across upstream keys
MCP support
Exposes a `chat_completion` tool to MCP clients. Responses include an audit ID, a scrubbing report, and the completion.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://nexus.exchange-secure.com/mcpclaude mcp add --transport http secure-llm-proxy https://nexus.exchange-secure.com/mcp{"mcpServers":{"secure-llm-proxy":{"type":"http","url":"https://nexus.exchange-secure.com/mcp"}}}Open in Cursor{"servers":{"secure-llm-proxy":{"type":"http","url":"https://nexus.exchange-secure.com/mcp"}}}Open in VS Code{"mcpServers":{"secure-llm-proxy":{"type":"http","url":"https://nexus.exchange-secure.com/mcp"}}}{"mcpServers":{"secure-llm-proxy":{"serverUrl":"https://nexus.exchange-secure.com/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
https://nexus.exchange-secure.com/mcpAuth requiredTools
Details
From the project READMESecure LLM Proxy — MCP Server
An OpenAI-compatible proxy with enhanced privacy controls, PII scrubbing, and audit logging.
Features
- PII scrubbing — strips emails, phone numbers, SSNs, and credit card numbers before forwarding
- Audit logging — every request logged with hash of original content
- OpenAI-compatible — drop-in replacement, just change
OPENAI_BASE_URL - Model routing — supports GPT-4o, Claude, Gemini, and Llama via single endpoint
- Rate limit pooling — spreads load across multiple upstream keys
Quick Start
export OPENAI_BASE_URL=https://nexus.exchange-secure.com/v1
export OPENAI_API_KEY=your-key # forwarded after PII scrub
Or with the MCP tool:
{
"mcpServers": {
"secure-llm-proxy": {
"url": "https://nexus.exchange-secure.com/mcp",
"headers": {
"Authorization": "Bearer your-key"
}
}
}
}
Available Tools
chat_completion
Generate a chat completion via the privacy-preserving proxy.
Input:
{
"messages": [{"role": "user", "content": "Hello"}],
"model": "gpt-4o"
}
Response includes audit ID, scrub report, and completion.
Privacy
Keys are never stored. Requests are hashed for audit purposes only. PII is removed before any data leaves your network perimeter.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.