@juspay/neurolink 9.68.17 → 9.68.19

package/dist/providers/openRouter.js

@@ -1,44 +1,30 @@
-import { createOpenRouter } from "@openrouter/ai-sdk-provider";
 import { AIProviderName } from "../constants/enums.js";
-import { BaseProvider } from "../core/baseProvider.js";
-import { DEFAULT_MAX_STEPS } from "../core/constants.js";
-import { streamAnalyticsCollector } from "../core/streamAnalytics.js";
-import { createProxyFetch } from "../proxy/proxyFetch.js";
 import { AuthenticationError, InvalidModelError, NetworkError, ProviderError, RateLimitError, } from "../types/index.js";
+import { createProxyFetch } from "../proxy/proxyFetch.js";
 import { isAbortError } from "../utils/errorHandling.js";
-import { emitToolEndFromStepFinish } from "../utils/toolEndEmitter.js";
 import { logger } from "../utils/logger.js";
-import { buildNoOutputSentinel, detectPostStreamNoOutput, stampNoOutputSpan, } from "../utils/noOutputSentinel.js";
+import { redactUrlCredentials } from "../utils/logSanitize.js";
 import { getProviderModel } from "../utils/providerConfig.js";
-import { composeAbortSignals, createTimeoutController, TimeoutError, } from "../utils/timeout.js";
-import { resolveToolChoice } from "../utils/toolChoice.js";
-import { NoOutputGeneratedError } from "../utils/generationErrors.js";
-import { Output, stepCountIs } from "../utils/tool.js";
-import { streamText } from "../utils/generation.js";
-// Constants
+import { TimeoutError } from "../utils/timeout.js";
+import { OpenAIChatCompletionsProvider } from "./openaiChatCompletionsBase.js";
+import { stripTrailingSlash } from "./openaiChatCompletionsClient.js";
+// OpenRouter's OpenAI-compatible gateway. `${baseURL}/chat/completions` and
+// `${baseURL}/models` both resolve correctly off this root.
+const OPENROUTER_DEFAULT_BASE_URL = "https://openrouter.ai/api/v1";
 const MODELS_DISCOVERY_TIMEOUT_MS = 5000; // 5 seconds for model discovery
-// Configuration helpers
-const getOpenRouterConfig = () => {
+const getOpenRouterApiKey = () => {
     const apiKey = process.env.OPENROUTER_API_KEY;
     if (!apiKey) {
         throw new Error("OPENROUTER_API_KEY environment variable is required. " +
             "Get your API key at https://openrouter.ai/keys");
     }
-    return {
-        apiKey,
-        referer: process.env.OPENROUTER_REFERER,
-        appName: process.env.OPENROUTER_APP_NAME,
-    };
+    return apiKey;
 };
 /**
  * Returns the default model name for OpenRouter.
  *
- * OpenRouter uses a 'provider/model' format for model names.
- * For example:
- *   - 'anthropic/claude-sonnet-4.5'
- *   - 'openai/gpt-4o'
- *   - 'google/gemini-2.5-flash'
- *   - 'meta-llama/llama-3-70b-instruct'
+ * OpenRouter uses a 'provider/model' format for model names (e.g.
+ * 'anthropic/claude-sonnet-4.5', 'openai/gpt-4o', 'google/gemini-2.5-flash').
  *
  * The previous default `anthropic/claude-3-5-sonnet` was retired by OpenRouter
  * in late 2025 and now returns "No endpoints found for model" for every
@@ -47,20 +33,34 @@ const getOpenRouterConfig = () => {
  * model. Must stay aligned with the registry default in
  * `src/lib/factories/providerRegistry.ts` and `PROVIDER_DEFAULTS` in
  * `src/lib/utils/modelChoices.ts`.
- *
- * You can override the default by setting the OPENROUTER_MODEL environment variable.
  */
 const getDefaultOpenRouterModel = () => {
     return getProviderModel("OPENROUTER_MODEL", "anthropic/claude-sonnet-4.5");
 };
 /**
- * OpenRouter Provider - BaseProvider Implementation
- * Provides access to 300+ models from 60+ providers via OpenRouter unified gateway
+ * OpenRouter Provider — direct HTTP, no AI SDK.
+ *
+ * OpenAI-compatible unified gateway to 300+ models from 60+ providers. All
+ * request/stream/tool-loop orchestration lives in
+ * `OpenAIChatCompletionsProvider`; this class declares configuration plus the
+ * OpenRouter-specific behaviour:
+ *
+ *   1. Attribution headers — optional `HTTP-Referer` / `X-Title` (from
+ *      `OPENROUTER_REFERER` / `OPENROUTER_APP_NAME`) are merged into every
+ *      request via `getAuthHeaders` so usage shows up on the openrouter.ai
+ *      activity dashboard.
+ *   2. Per-model tool gating — OpenRouter proxies many models with varying
+ *      tool support, so `supportsTools()` consults a cached capability set
+ *      (populated by `cacheModelCapabilities()`) and falls back to a
+ *      conservative known-capable pattern list.
+ *   3. Dynamic model discovery — `getAvailableModels()` fetches the live
+ *      `/models` list (10-minute cache) with a hardcoded fallback.
+ *
+ * @see https://openrouter.ai/docs
  */
-export class OpenRouterProvider extends BaseProvider {
-    model;
-    openRouterClient;
-    config;
+export class OpenRouterProvider extends OpenAIChatCompletionsProvider {
+    referer;
+    appName;
     // Cache for available models to avoid repeated API calls
     static modelsCache = [];
     static modelsCacheTime = 0;
@@ -69,53 +69,40 @@ export class OpenRouterProvider extends BaseProvider {
     static toolCapableModels = new Set();
     static capabilitiesCached = false;
     constructor(modelName, sdk, _region, credentials) {
-        super(modelName, AIProviderName.OPENROUTER, sdk);
-        // Build config: prefer credentials over env vars to avoid throwing when env vars are absent
-        if (credentials?.apiKey) {
-            this.config = {
-                apiKey: credentials.apiKey,
-                referer: process.env.OPENROUTER_REFERER,
-                appName: process.env.OPENROUTER_APP_NAME,
-            };
-        }
-        else {
-            this.config = getOpenRouterConfig(); // throws if OPENROUTER_API_KEY missing
-        }
-        const config = this.config;
-        // Build headers for attribution on openrouter.ai/activity dashboard
-        const headers = {};
-        if (config.referer) {
-            headers["HTTP-Referer"] = config.referer;
-        }
-        if (config.appName) {
-            headers["X-Title"] = config.appName;
-        }
-        // Create OpenRouter client with optional attribution headers
-        this.openRouterClient = createOpenRouter({
-            apiKey: config.apiKey,
-            ...(credentials?.baseURL ? { baseURL: credentials.baseURL } : {}),
-            ...(Object.keys(headers).length > 0 && { headers }),
-        });
-        // Initialize model with OpenRouter client
-        // OpenRouterChatLanguageModel implements LanguageModelV3 which is part of the LanguageModel union
-        this.model = this.openRouterClient(this.modelName || getDefaultOpenRouterModel());
+        // Trim the override before applying precedence. A blank/whitespace
+        // `credentials.apiKey` must NOT bypass `getOpenRouterApiKey()` — that
+        // would build a client with an unusable bearer token and fail at request
+        // time with a confusing 401 instead of at construction time.
+        const overrideApiKey = credentials?.apiKey?.trim();
+        const apiKey = overrideApiKey && overrideApiKey.length > 0
+            ? overrideApiKey
+            : getOpenRouterApiKey();
+        // Treat blank/whitespace overrides as unset so an empty
+        // `credentials.baseURL` or `OPENROUTER_BASE_URL=` cannot silently override
+        // the default with "" (mirrors the apiKey precedence above).
+        const baseURL = credentials?.baseURL?.trim() ||
+            process.env.OPENROUTER_BASE_URL?.trim() ||
+            OPENROUTER_DEFAULT_BASE_URL;
+        super(AIProviderName.OPENROUTER, modelName, sdk, { baseURL, apiKey });
+        // Trim attribution values so whitespace-only env vars are not sent as
+        // empty HTTP-Referer / X-Title headers.
+        this.referer = process.env.OPENROUTER_REFERER?.trim() || undefined;
+        this.appName = process.env.OPENROUTER_APP_NAME?.trim() || undefined;
         logger.debug("OpenRouter Provider initialized", {
             modelName: this.modelName,
-            provider: this.providerName,
+            providerName: this.providerName,
+            baseURL: redactUrlCredentials(this.config.baseURL),
         });
     }
+    // ===========================================================================
+    // Abstract hooks (required)
+    // ===========================================================================
     getProviderName() {
         return AIProviderName.OPENROUTER;
     }
     getDefaultModel() {
         return getDefaultOpenRouterModel();
     }
-    /**
-     * Returns the Vercel AI SDK model instance for OpenRouter
-     */
-    getAISDKModel() {
-        return this.model;
-    }
     formatProviderError(error) {
         if (error instanceof TimeoutError) {
             return new NetworkError(`Request timed out: ${error.message}`, "openrouter");
@@ -150,9 +137,9 @@ export class OpenRouterProvider extends BaseProvider {
             if (errorRecord.message.includes("insufficient_credits")) {
                 return new ProviderError("Insufficient OpenRouter credits. Add credits at https://openrouter.ai/credits", "openrouter");
             }
-            // "No endpoints found" — model temporarily unavailable or unsupported parameters
-            // This is distinct from tool errors: it can happen on any request when the
-            // model has no available providers on OpenRouter (e.g., free-tier model down).
+            // "No endpoints found" — model temporarily unavailable or unsupported
+            // parameters. Distinct from tool errors: it can happen on any request
+            // when the model has no available providers on OpenRouter.
             if (errorRecord.message.includes("No endpoints found")) {
                 return new InvalidModelError(`No endpoints found for model '${this.modelName}' on OpenRouter. ` +
                     "The model may be temporarily unavailable or does not support the requested parameters. " +
@@ -175,9 +162,31 @@ export class OpenRouterProvider extends BaseProvider {
         }
         return new ProviderError(`OpenRouter error: ${errorRecord?.message || "Unknown error"}`, "openrouter");
     }
+    // ===========================================================================
+    // Optional hooks — provider-specific quirks
+    // ===========================================================================
+    getFallbackModelName() {
+        return getDefaultOpenRouterModel();
+    }
+    /**
+     * Attribution headers are merged into every request alongside the bearer
+     * token so OpenRouter can attribute usage on its activity dashboard.
+     */
+    getAuthHeaders() {
+        const headers = { ...super.getAuthHeaders() };
+        if (this.referer) {
+            headers["HTTP-Referer"] = this.referer;
+        }
+        if (this.appName) {
+            headers["X-Title"] = this.appName;
+        }
+        return headers;
+    }
     /**
-     * OpenRouter supports tools for compatible models
-     * Checks cached model capabilities or uses known patterns as fallback
+     * OpenRouter proxies models with varying tool support. Use cached
+     * capabilities when available (populated by `cacheModelCapabilities()`),
+     * otherwise fall back to a conservative known-capable pattern list and
+     * disable tools for unknown models.
      */
     supportsTools() {
         const modelName = this.modelName || getDefaultOpenRouterModel();
@@ -222,220 +231,19 @@ export class OpenRouterProvider extends BaseProvider {
         });
         return false;
     }
+    // ===========================================================================
+    // Model discovery (OpenRouter /models endpoint)
+    // ===========================================================================
     /**
-     * Provider-specific streaming implementation
-     * Note: This is only used when tools are disabled
+     * Models/capabilities endpoint, derived from the configured `baseURL` so a
+     * custom OpenRouter-compatible gateway is honoured for discovery too.
      */
-    async executeStream(options, analysisSchema) {
-        this.validateStreamOptions(options);
-        const startTime = Date.now();
-        let chunkCount = 0; // Track chunk count for debugging
-        // Reviewer follow-up: capture upstream provider errors via onError so
-        // the post-stream NoOutput detect can propagate the *real* cause
-        // (e.g. content_filter, provider crash) into the sentinel's
-        // providerError / modelResponseRaw instead of the AI SDK's generic
-        // "No output generated" message.
-        let capturedProviderError;
-        const timeout = this.getTimeout(options);
-        const timeoutController = createTimeoutController(timeout, this.providerName, "stream");
-        try {
-            // Build message array from options with multimodal support
-            // Using protected helper from BaseProvider to eliminate code duplication
-            const messages = await this.buildMessagesForStream(options);
-            const model = await this.getAISDKModelWithMiddleware(options);
-            // Get all available tools (direct + MCP + external) for streaming
-            // BaseProvider.stream() pre-merges base tools + external tools into options.tools
-            const shouldUseTools = !options.disableTools && this.supportsTools();
-            const tools = shouldUseTools
-                ? options.tools || (await this.getAllTools())
-                : {};
-            logger.debug(`OpenRouter: Tools for streaming`, {
-                shouldUseTools,
-                toolCount: Object.keys(tools).length,
-                toolNames: Object.keys(tools),
-            });
-            // Build complete stream options with proper typing
-            // Note: maxRetries set to 0 for OpenRouter free tier to prevent SDK's quick retries
-            // from consuming rate limits. Our test suite handles retries with appropriate delays.
-            let streamOptions = {
-                model: model,
-                messages: messages,
-                temperature: options.temperature,
-                maxRetries: 0, // Disable SDK retries - let caller handle rate limit retries with delays
-                // AI SDK v6 renamed `maxTokens` to `maxOutputTokens` — using the old
-                // name here is a silent no-op, so OpenRouter sees no cap and applies
-                // the model's full output max (typically 64K+ tokens) to its pre-bill
-                // affordability check. That trips "This request requires more credits"
-                // even on cheap models when the account balance is low.
-                ...(options.maxTokens && { maxOutputTokens: options.maxTokens }),
-                ...(shouldUseTools &&
-                    Object.keys(tools).length > 0 && {
-                    tools,
-                    toolChoice: resolveToolChoice(options, tools, shouldUseTools),
-                    stopWhen: stepCountIs(options.maxSteps || DEFAULT_MAX_STEPS),
-                }),
-                abortSignal: composeAbortSignals(options.abortSignal, timeoutController?.controller.signal),
-                experimental_telemetry: this.telemetryHandler.getTelemetryConfig(options),
-                experimental_repairToolCall: this.getToolCallRepairFn(options),
-                onError: (event) => {
-                    const error = event.error;
-                    const errorMessage = error instanceof Error ? error.message : String(error);
-                    // Reviewer follow-up: propagate the captured error to the
-                    // post-stream NoOutput sentinel so telemetry sees the real
-                    // provider cause instead of "No output generated".
-                    capturedProviderError = error;
-                    logger.error(`OpenRouter: Stream error`, {
-                        provider: this.providerName,
-                        modelName: this.modelName,
-                        error: errorMessage,
-                        chunkCount,
-                    });
-                },
-                onFinish: (event) => {
-                    logger.debug(`OpenRouter: Stream finished`, {
-                        finishReason: event.finishReason,
-                        totalChunks: chunkCount,
-                    });
-                },
-                onChunk: () => {
-                    chunkCount++;
-                },
-                onStepFinish: ({ toolCalls, toolResults }) => {
-                    emitToolEndFromStepFinish(this.neurolink?.getEventEmitter(), toolResults);
-                    logger.info("Tool execution completed", {
-                        toolCallCount: toolCalls?.length || 0,
-                        toolResultCount: toolResults?.length || 0,
-                        toolNames: toolCalls?.map((tc) => tc.toolName),
-                    });
-                    this.handleToolExecutionStorage(toolCalls, toolResults, options, new Date()).catch((error) => {
-                        logger.warn("OpenRouterProvider: Failed to store tool executions", {
-                            provider: this.providerName,
-                            error: error instanceof Error ? error.message : String(error),
-                        });
-                    });
-                },
-            };
-            // Add analysisSchema support if provided
-            if (analysisSchema) {
-                try {
-                    streamOptions = {
-                        ...streamOptions,
-                        experimental_output: Output.object({
-                            schema: analysisSchema,
-                        }),
-                    };
-                }
-                catch (error) {
-                    logger.warn("Schema application failed, continuing without schema", {
-                        error: String(error),
-                    });
-                }
-            }
-            const result = await streamText(streamOptions);
-            // Guard against NoOutputGeneratedError becoming an unhandled rejection.
-            Promise.resolve(result.text)
-                .catch((err) => {
-                logger.debug("Stream text promise rejected (expected for empty streams)", {
-                    error: err instanceof Error ? err.message : String(err),
-                });
-            })
-                .finally(() => timeoutController?.cleanup());
-            // Transform stream to content object stream using fullStream (handles both text and tool calls)
-            const transformedStream = (async function* () {
-                // Reviewer follow-up: gate the post-stream NoOutput detect on
-                // *content yielded*, not raw chunk count. AI SDK fullStream emits
-                // control events ({ type: "start" }, "step-start", etc.) before
-                // any text-delta — those incremented `chunkCount` and made the
-                // post-stream check dead even when zero text was produced.
-                let contentYielded = 0;
-                try {
-                    // Try fullStream first (handles both text and tool calls), fallback to textStream
-                    const streamToUse = result.fullStream || result.textStream;
-                    for await (const chunk of streamToUse) {
-                        // Handle different chunk types from fullStream
-                        if (chunk && typeof chunk === "object") {
-                            // Check for error chunks first (critical error handling)
-                            if ("type" in chunk && chunk.type === "error") {
-                                const errorChunk = chunk;
-                                logger.error(`OpenRouter: Error chunk received:`, {
-                                    errorType: errorChunk.type,
-                                    errorDetails: errorChunk.error,
-                                });
-                                throw new Error(`OpenRouter streaming error: ${errorChunk.error?.message ||
-                                    "Unknown error"}`);
-                            }
-                            if ("textDelta" in chunk) {
-                                // Text delta from fullStream
-                                const textDelta = chunk.textDelta;
-                                if (textDelta) {
-                                    contentYielded++;
-                                    yield { content: textDelta };
-                                }
-                            }
-                            else if ("type" in chunk &&
-                                chunk.type === "tool-call" &&
-                                "toolCallId" in chunk) {
-                                // Tool call event - log for debugging
-                                const toolCallId = String(chunk.toolCallId);
-                                const toolName = "toolName" in chunk ? String(chunk.toolName) : "unknown";
-                                logger.debug("OpenRouter: Tool call", {
-                                    toolCallId,
-                                    toolName,
-                                });
-                            }
-                        }
-                        else if (typeof chunk === "string") {
-                            // Direct string chunk from textStream fallback
-                            contentYielded++;
-                            yield { content: chunk };
-                        }
-                    }
-                }
-                catch (streamError) {
-                    if (NoOutputGeneratedError.isInstance(streamError)) {
-                        logger.warn("OpenRouter: Stream produced no output (NoOutputGeneratedError) — caught from textStream");
-                        const sentinel = await buildNoOutputSentinel(streamError, result, capturedProviderError);
-                        stampNoOutputSpan(sentinel);
-                        yield sentinel;
-                        return;
-                    }
-                    throw streamError;
-                }
-                // Curator P3-6 (round-2 fix): production trigger comes through
-                // result.finishReason rejection, not textStream throws.
-                if (contentYielded === 0) {
-                    const detected = await detectPostStreamNoOutput(result, capturedProviderError);
-                    if (detected) {
-                        logger.warn("OpenRouter: Stream produced no output (NoOutputGeneratedError) — caught from finishReason rejection");
-                        stampNoOutputSpan(detected.sentinel);
-                        yield detected.sentinel;
-                    }
-                }
-            })();
-            // Create analytics promise that resolves after stream completion
-            const analyticsPromise = streamAnalyticsCollector.createAnalytics(this.providerName, this.modelName, result, Date.now() - startTime, {
-                requestId: `openrouter-stream-${Date.now()}`,
-                streamingMode: true,
-            });
-            return {
-                stream: transformedStream,
-                provider: this.providerName,
-                model: this.modelName,
-                analytics: analyticsPromise,
-                metadata: {
-                    startTime,
-                    streamId: `openrouter-${Date.now()}`,
-                },
-            };
-        }
-        catch (error) {
-            timeoutController?.cleanup();
-            throw this.handleProviderError(error);
-        }
+    getModelsUrl() {
+        return `${stripTrailingSlash(this.config.baseURL)}/models`;
     }
     /**
-     * Get available models from OpenRouter API
-     * Dynamically fetches from /api/v1/models endpoint with caching and fallback
+     * Get available models from the OpenRouter `/models` endpoint, with a
+     * 10-minute cache and a hardcoded fallback when the fetch fails.
      */
     async getAvailableModels() {
         const functionTag = "OpenRouterProvider.getAvailableModels";
@@ -454,7 +262,6 @@ export class OpenRouterProvider extends BaseProvider {
         try {
             const dynamicModels = await this.fetchModelsFromAPI();
             if (dynamicModels.length > 0) {
-                // Cache successful result
                 OpenRouterProvider.modelsCache = dynamicModels;
                 OpenRouterProvider.modelsCacheTime = now;
                 logger.debug(`[${functionTag}] Successfully fetched models from API`, {
@@ -496,22 +303,21 @@ export class OpenRouterProvider extends BaseProvider {
         return fallbackModels;
     }
     /**
-     * Fetch available models from OpenRouter API /api/v1/models endpoint
+     * Fetch available models from the OpenRouter `/models` endpoint.
      * @private
      */
     async fetchModelsFromAPI() {
         const functionTag = "OpenRouterProvider.fetchModelsFromAPI";
-        const config = this.config;
-        const modelsUrl = "https://openrouter.ai/api/v1/models";
         const controller = new AbortController();
         const timeoutId = setTimeout(() => controller.abort(), MODELS_DISCOVERY_TIMEOUT_MS);
         try {
+            const modelsUrl = this.getModelsUrl();
             logger.debug(`[${functionTag}] Fetching models from ${modelsUrl}`);
             const proxyFetch = createProxyFetch();
             const response = await proxyFetch(modelsUrl, {
                 method: "GET",
                 headers: {
-                    Authorization: `Bearer ${config.apiKey}`,
+                    ...this.getAuthHeaders(),
                     "Content-Type": "application/json",
                 },
                 signal: controller.signal,
@@ -544,7 +350,7 @@ export class OpenRouterProvider extends BaseProvider {
         }
     }
     /**
-     * Type guard to validate the models API response structure
+     * Type guard to validate the models API response structure.
      * @private
      */
     isValidModelsResponse(data) {
@@ -554,8 +360,8 @@ export class OpenRouterProvider extends BaseProvider {
             Array.isArray(data.data));
     }
     /**
-     * Fetch and cache model capabilities from OpenRouter API
-     * Call this to enable accurate tool support detection
+     * Fetch and cache model capabilities from the OpenRouter `/models` endpoint.
+     * Call this to enable accurate per-model tool support detection.
      */
     async cacheModelCapabilities() {
         const functionTag = "OpenRouterProvider.cacheModelCapabilities";
@@ -563,15 +369,13 @@ export class OpenRouterProvider extends BaseProvider {
             return; // Already cached
         }
         try {
-            const config = this.config;
-            const modelsUrl = "https://openrouter.ai/api/v1/models";
             const controller = new AbortController();
             const timeoutId = setTimeout(() => controller.abort(), MODELS_DISCOVERY_TIMEOUT_MS);
             const proxyFetch = createProxyFetch();
-            const response = await proxyFetch(modelsUrl, {
+            const response = await proxyFetch(this.getModelsUrl(), {
                 method: "GET",
                 headers: {
-                    Authorization: `Bearer ${config.apiKey}`,
+                    ...this.getAuthHeaders(),
                     "Content-Type": "application/json",
                 },
                 signal: controller.signal,

package/dist/types/providers.d.ts

@@ -129,6 +129,7 @@ export type NeurolinkCredentials = {
     };
     mistral?: {
         apiKey?: string;
+        baseURL?: string;
     };
     huggingFace?: {
         apiKey?: string;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@juspay/neurolink",
-  "version": "9.68.17",
+  "version": "9.68.19",
   "packageManager": "pnpm@10.15.1",
   "description": "Universal AI Development Platform with working MCP integration, multi-provider support, voice (TTS/STT/realtime), and professional CLI. 58+ external MCP servers discoverable, multimodal file processing, RAG pipelines. Build, test, and deploy AI applications with 21+ providers: OpenAI, Anthropic, Google AI Studio, Google Vertex, AWS Bedrock, Azure OpenAI, Mistral, LiteLLM, SageMaker, Hugging Face, Ollama, OpenAI-compatible, OpenRouter, DeepSeek, NVIDIA NIM, LM Studio, llama.cpp, plus voice (OpenAI TTS, ElevenLabs, Deepgram, Azure Speech).",
   "author": {