System Status

Some providers are degraded. Automatic failover routes around them where another provider can serve the model.

Live checks, refreshed every minute · last checked

TrivaNexus gateway

API and database responding

Operational

OpenRouter

Gemini, Grok, Mistral, Qwen, DeepSeek, Llama and all openrouter/* models

DownOut of credit — requests will fail272 ms

Hugging Face

huggingface/* models, and fallback for Llama, Qwen and DeepSeek

OperationalResponding normally191 ms

Kimi (Moonshot AI)

kimi/* models

DownOut of credit — requests will fail272 ms

OpenAI

openai/* models

Not configuredNot configured on this gateway—

Anthropic

anthropic/* models

Not configuredNot configured on this gateway—

Local server

local/* models

Not configuredNot configured on this gateway—

Checks use each provider's account or model-list endpoint, so they confirm the API is reachable, the key is accepted and — where the provider reports it — that the account has credit. They don't send model requests, so an individual model can still fail while its provider shows as operational. Machine-readable: /api/status.