Hey Product Hunt! 👋 We built SentinelGateway out of frustration with single-provider fragility in production. Whenever OpenAI throws a 503 or Anthropic hits a 529 overload, client apps crash. On top of that, sending raw customer prompts to third-party APIs without sanitization is a compliance liability. Rather than forcing teams to install heavy wrapper SDKs or deal with high proxy overhead, we engineered SentinelGateway in Go: Zero SDK changes: Swap your base_url in standard OpenAI clients. Resilience: Auto-failover across Claude, GPT-4o, Groq, and Gemini in <25ms. Privacy first: Zero-retention, in-flight PII redaction in memory. Speed & cost: Atomic Redis semantic caching to slash repeat token costs. We are live with a free tier so you can test failover routing in your staging environment right away. Would love your feedback on the routing performance, proxy latency, or features you’d like to see next!

SentinelGateway

OpenAI-compatible Go proxy with auto-failover & PII redact

OpenAI-compatible Go proxy with auto-failover & PII redact