@wengine-ai/llms 2.3.238 → 2.3.239

package/dist/services/active-probe.d.ts

@@ -13,11 +13,35 @@ export interface ActiveProbeConfig {
     quotaProbeIntervalMinutes?: number;
     /** Timeout in milliseconds for probe requests (default: 15000) */
     probeTimeoutMs?: number;
+    /** Latency threshold in milliseconds for slow-network warnings (default: 3000) */
+    slowThresholdMs?: number;
     /** Initial delay before starting first probe (default: 5000) */
     initialDelayMs?: number;
     /** Providers to exclude from active probing (e.g., providers without models endpoint) */
     excludeProviders?: string[];
 }
+export type ProviderProbeStatus = 'healthy' | 'slow' | 'error' | 'timeout';
+export type ProviderProbeErrorKind = 'timeout' | 'network' | 'http';
+export type ProviderProbeSource = 'health' | 'manual' | 'rate-limit-headers';
+export interface ProviderProbeTelemetry {
+    provider: string;
+    latencyMs: number;
+    status: ProviderProbeStatus;
+    isSlow: boolean;
+    errorKind?: ProviderProbeErrorKind;
+    errorMessage?: string;
+    lastProbeAt: number;
+    lastSuccessAt?: number;
+    source: ProviderProbeSource;
+}
+export interface ManualProbeResult {
+    success: boolean;
+    latencyMs: number;
+    status: ProviderProbeStatus;
+    isSlow: boolean;
+    errorKind?: ProviderProbeErrorKind;
+    error?: string;
+}
 /**
  * Active probe scheduler
  */
@@ -33,7 +57,23 @@ export declare class ActiveProbeService {
     private logger?;
     private getConfig?;
     private running;
+    private probeTelemetry;
     constructor(getProviders: () => LLMProvider[], config?: ActiveProbeConfig, resolveProxyUrl?: (provider: LLMProvider) => string | undefined, logger?: any, getConfig?: (key: string) => any);
+    /**
+     * Record the latest reachability probe result for a provider.
+     * Only /models-style reachability probes (health, manual, rate-limit header
+     * probes) should be recorded here; quota adapters, wake-up calls and
+     * rate-limit chat pings hit different endpoints and must not overwrite it.
+     */
+    private recordProbe;
+    /**
+     * Read-only snapshot of the latest probe telemetry for all providers.
+     */
+    getProbeTelemetry(): ProviderProbeTelemetry[];
+    /**
+     * Read-only snapshot of the latest probe telemetry for one provider.
+     */
+    getProviderProbeTelemetry(providerName: string): ProviderProbeTelemetry | undefined;
     /**
      * Start the active probe loop
      */
@@ -76,7 +116,7 @@ export declare class ActiveProbeService {
     /**
      * Trigger a manual probe for a specific provider
      */
-    probeProviderManually(providerName: string): Promise<boolean>;
+    probeProviderManually(providerName: string): Promise<ManualProbeResult>;
 }
 /**
  * Initialize and get the active probe service

package/dist/utils/router.d.ts

@@ -29,6 +29,11 @@ interface MessageParam {
     content: string | ContentBlockParam[];
 }
 export declare const calculateTokenCount: (messages: MessageParam[], system: any, tools: Tool[]) => number;
+export declare function extractModelFamily(modelName: string): {
+    family: string | null;
+    extended: boolean;
+    isCcrAlias: boolean;
+};
 /**
  * Look up a provider and model by name in the current provider list.
  * Returns canonical (case-matched) values or null when the provider is
@@ -47,6 +52,7 @@ export declare function diagnoseResolutionFailure(configuredModel: string, provi
     reason: string;
     code: string;
 };
+export declare function modelSupportsImages(modelName: string): boolean;
 interface RouterFamilyConfig {
     default: string;
     background?: string;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@wengine-ai/llms",
-  "version": "2.3.238",
+  "version": "2.3.239",
   "description": "A universal LLM API transformation server",
   "main": "dist/cjs/server.cjs",
   "module": "dist/esm/server.mjs",
@@ -52,7 +52,7 @@
     "uuid": "^11.1.0",
     "@fastify/multipart": "^9.0.0",
     "@fastify/static": "^8.2.0",
-    "@wengine-ai/claude-code-router-shared": "^2.3.238",
+    "@wengine-ai/claude-code-router-shared": "^2.3.239",
     "adm-zip": "^0.5.16",
     "better-sqlite3": "^12.10.0",
     "rotating-file-stream": "^3.2.7"

package/dist/ccr/codex-accounts.d.ts

@@ -1,9 +0,0 @@
-export declare function switchCodexAccountBeforeUsageLimit(): Promise<void>;
-export declare function switchCodexAccountAfterRateLimit(reason?: string): Promise<void>;
-/** Periodically refresh managed Codex account tokens, active or not, so refresh_tokens never go stale. */
-export declare function startCodexTokenRefreshScheduler(): void;
-export declare function isRateLimitMessage(statusCode: number, message?: string): boolean;
-export declare function getCurrentCodexAccountForRequest(): Promise<{
-    id?: string;
-    email?: string;
-}>;

package/dist/ccr/codex-usage-cache.d.ts

@@ -1,2 +0,0 @@
-import { type CodexAccountsResult } from "@wengine-ai/claude-code-router-shared";
-export declare function computeCodexAccountUsage(config: Record<string, any>): Promise<CodexAccountsResult>;