A rolled-up view of everything below — real checks, condensed into one alert list.
Live availability and latency for every service on this platform, polled every 30 seconds over the internal Docker network.
Request volume, latency, error rate and model mix, computed from real Langfuse trace data.
Real LLM spend from Langfuse trace data. Cloud hosting, domain and other infrastructure costs are not tracked here yet.
Certificate validity, guardrail activity, auth activity and a checklist of what is actually wired up on this platform.
Recent conversations across BuyBuddy and every other app on this platform, pulled live from Langfuse.
This sends your message through the real shared LiteLLM proxy — the same gateway every app on this platform uses — with OpenAI's Moderation guardrail active. Nothing here is scripted.