Configuration → Security is where you control what the gateway is allowed to do beyond plain LLM inference: which MCP servers it exposes to clients, which content/intent rules it enforces on every request, and which on-gateway SLM (small language model) classifiers run as additional guards.The page has three tabs: MCP, Guardrails, and SLM Guardrails.
Register the Model Context Protocol servers your gateway should make available to clients. Once registered, MCP servers are scoped per API key on the API Keys MCP tab.
Open Configuration → Security from the dashboard sidebar and click the MCP tab.
2
Register a server
Click Register MCP Server. Pick a Template to autofill a known integration (filesystem, github, etc.) or switch to Manual.
Register MCP Server
3
Pick the transport
Choose STDIO for processes the gateway spawns, or HTTP / SSE for remote MCP servers. Fill the matching fields.
4
Set authentication
Select an auth scheme and provide the secret. Use None only for local STDIO servers on the gateway host.
5
Save and scope to keys
The server appears in the list with status connected once the gateway can reach it. Open API Keys, edit a key, and use the MCP tab to scope the key to specific servers and tools.
Rule-based content controls applied to every request that flows through the gateway. Use these to redact PII, block prompt injection patterns, enforce keyword filters, or just log policy hits.
Block — reject the request, Warn — Allow but flag, Redact — replace matched content, Log — Record only.
Description
Free-form notes.
Apply to MCP
Toggle. When on, the rule also runs against MCP traffic on this gateway.
Keywords to block
One per line — for keyword-style categories.
Regex patterns (optional)
One per line — e.g. \b\d{3}-\d{2}-\d{4}\b.
Ignore case
Toggle for keyword/regex matches.
Rules are listed with Name, Category, Type (input / output / both), Action, Priority, Triggers, and Enabled. A built-in Test dialog lets you paste sample text and see which rules fire.
On-gateway small language models that classify, score, detect, or rewrite content. Use these when keyword/regex rules aren’t expressive enough — toxicity classifiers, jailbreak detectors, intent scorers.
