DOCS / MANAGEMENT API / PROVIDERS & MODELS
Providers & Models
Provider health
http
Authorization: Bearer lattice_sk_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxjson
type ProviderAdapter = {
id: ProviderId;
createChatCompletion(
request: ChatCompletionRequest,
ctx: ProviderRequestContext,
): Promise<ChatCompletionResponse>;
streamChatCompletion(
request: ChatCompletionRequest,
ctx: ProviderRequestContext,
): AsyncGenerator<ChatCompletionChunk, void, void>;
};
type ProviderRequestContext = {
apiKey?: string;
providerModel: string;
signal: AbortSignal;
};circuitState is 'closed' | 'open' | 'half-open' — read live from the circuit breaker, synced across gateway instances every 5 seconds. averageLatencyMs is null until at least one successful request has been recorded for that model. This is the endpoint the dashboard's provider health board polls.
Model registry
http
{
"id": "chatcmpl-...",
"object": "chat.completion",
"created": 1755878400,
"model": "gpt-4o",
"choices": [
{
"index": 0,
"message": { "role": "assistant", "content": "..." },
"finish_reason": "stop"
}
],
"usage": {
"prompt_tokens": 42,
"completion_tokens": 118,
"total_tokens": 160
}
}json
data: {"id":"chatcmpl-...","object":"chat.completion.chunk","created":1755878400,"model":"gpt-4o","choices":[{"index":0,"delta":{"role":"assistant","content":""},"finish_reason":null}]}
data: {"id":"chatcmpl-...","object":"chat.completion.chunk","created":1755878400,"model":"gpt-4o","choices":[{"index":0,"delta":{"content":"Sure"},"finish_reason":null}]}
data: {"id":"chatcmpl-...","object":"chat.completion.chunk","created":1755878400,"model":"gpt-4o","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":42,"completion_tokens":118,"total_tokens":160}}
data: [DONE] This is the fully resolved, static registry from apps/gateway/server/registry/models.config.ts — read-only, and identical regardless of which application queries it (there's no per-application filtering here; allowlists are enforced at request time on chat completions, not reflected in this listing). Changing the registry requires a code change and redeploy — see Model Registry & Routing.