Recipes

End-to-end walkthroughs for common jobs. Each one lists what you need, the exact settings, how to check the result, and the limits that apply.

recipe	for
Protect a LangGraph support bot	a LangChain or LangGraph agent with internal and external tools
Put the gateway in front of a GitHub MCP server	an MCP client using a third party MCP server, no code changes
Roll out in audit mode first, then enforce	adding Taintward to an agent already in production
Send traces to SigNoz, Grafana Tempo or Jaeger	wiring the collector to your trace backend
Generate the monthly report for a customer security review	answering "do you send our data to LLM providers?" with evidence

All examples use fake data. Never test with real personal data or real keys.