@openclaw/groq-provider 2026.7.2-beta.2 → 2026.7.2-beta.4

package/dist/index.js

@@ -1,8 +1,254 @@
 import { groqMediaUnderstandingProvider } from "./media-understanding-provider.js";
+import { createAssistantMessageEventStream, streamSimple } from "openclaw/plugin-sdk/llm";
 import { definePluginEntry } from "openclaw/plugin-sdk/plugin-entry";
 import { createProviderApiKeyAuthMethod } from "openclaw/plugin-sdk/provider-auth-api-key";
+import { buildOpenAICompatibleProviderCatalog } from "openclaw/plugin-sdk/provider-catalog-live-runtime";
+import { buildManifestModelProviderConfig } from "openclaw/plugin-sdk/provider-catalog-shared";
+//#region extensions/groq/openclaw.plugin.json
+var modelCatalog = {
+	"providers": { "groq": {
+		"baseUrl": "https://api.groq.com/openai/v1",
+		"api": "openai-completions",
+		"models": [
+			{
+				"id": "groq/compound",
+				"name": "Compound",
+				"reasoning": true,
+				"input": ["text"],
+				"contextWindow": 131072,
+				"maxTokens": 8192,
+				"cost": {
+					"input": 0,
+					"output": 0,
+					"cacheRead": 0,
+					"cacheWrite": 0
+				}
+			},
+			{
+				"id": "groq/compound-mini",
+				"name": "Compound Mini",
+				"reasoning": true,
+				"input": ["text"],
+				"contextWindow": 131072,
+				"maxTokens": 8192,
+				"cost": {
+					"input": 0,
+					"output": 0,
+					"cacheRead": 0,
+					"cacheWrite": 0
+				}
+			},
+			{
+				"id": "llama-3.1-8b-instant",
+				"name": "Llama 3.1 8B Instant",
+				"reasoning": false,
+				"input": ["text"],
+				"contextWindow": 131072,
+				"maxTokens": 131072,
+				"cost": {
+					"input": .05,
+					"output": .08,
+					"cacheRead": 0,
+					"cacheWrite": 0
+				}
+			},
+			{
+				"id": "llama-3.3-70b-versatile",
+				"name": "Llama 3.3 70B Versatile",
+				"reasoning": false,
+				"input": ["text"],
+				"contextWindow": 131072,
+				"maxTokens": 32768,
+				"cost": {
+					"input": .59,
+					"output": .79,
+					"cacheRead": 0,
+					"cacheWrite": 0
+				}
+			},
+			{
+				"id": "meta-llama/llama-4-scout-17b-16e-instruct",
+				"name": "Llama 4 Scout 17B",
+				"reasoning": false,
+				"input": ["text", "image"],
+				"mediaInput": { "image": {
+					"maxPixels": 33177600,
+					"preferredSidePx": 2048,
+					"tokenMode": "provider"
+				} },
+				"contextWindow": 131072,
+				"maxTokens": 8192,
+				"cost": {
+					"input": .11,
+					"output": .34,
+					"cacheRead": 0,
+					"cacheWrite": 0
+				}
+			},
+			{
+				"id": "openai/gpt-oss-120b",
+				"name": "GPT OSS 120B",
+				"reasoning": true,
+				"input": ["text"],
+				"contextWindow": 131072,
+				"maxTokens": 65536,
+				"cost": {
+					"input": .15,
+					"output": .6,
+					"cacheRead": 0,
+					"cacheWrite": 0
+				}
+			},
+			{
+				"id": "openai/gpt-oss-20b",
+				"name": "GPT OSS 20B",
+				"reasoning": true,
+				"input": ["text"],
+				"contextWindow": 131072,
+				"maxTokens": 65536,
+				"cost": {
+					"input": .075,
+					"output": .3,
+					"cacheRead": 0,
+					"cacheWrite": 0
+				}
+			},
+			{
+				"id": "openai/gpt-oss-safeguard-20b",
+				"name": "Safety GPT OSS 20B",
+				"reasoning": true,
+				"input": ["text"],
+				"contextWindow": 131072,
+				"maxTokens": 65536,
+				"cost": {
+					"input": .075,
+					"output": .3,
+					"cacheRead": .037,
+					"cacheWrite": 0
+				}
+			},
+			{
+				"id": "qwen/qwen3-32b",
+				"name": "Qwen3 32B",
+				"reasoning": true,
+				"input": ["text"],
+				"contextWindow": 131072,
+				"maxTokens": 40960,
+				"cost": {
+					"input": .29,
+					"output": .59,
+					"cacheRead": 0,
+					"cacheWrite": 0
+				}
+			}
+		]
+	} },
+	"discovery": { "groq": "refreshable" }
+};
+//#endregion
 //#region extensions/groq/index.ts
 const GROQ_DEFAULT_MODEL_REF = "groq/llama-3.3-70b-versatile";
+const GROQ_DEFAULT_MODEL_ID = "llama-3.3-70b-versatile";
+const GROQ_FALLBACK_MAX_TOKENS = 1024;
+function buildGroqCatalogProvider() {
+	return buildManifestModelProviderConfig({
+		providerId: "groq",
+		catalog: modelCatalog.providers.groq
+	});
+}
+function hasWireMaxTokens(value) {
+	if (typeof value !== "object" || value === null || Array.isArray(value)) return false;
+	const record = value;
+	return record.max_completion_tokens !== void 0 || record.max_tokens !== void 0;
+}
+function hasExplicitMaxTokens(extraParams) {
+	return extraParams?.maxTokens !== void 0 || hasWireMaxTokens(extraParams) || hasWireMaxTokens(extraParams?.extra_body) || hasWireMaxTokens(extraParams?.extraBody);
+}
+function isGroqTpmRequestTooLargeEvent(event) {
+	if (event.type !== "error") return false;
+	const message = event.error.errorMessage?.toLowerCase() ?? "";
+	return message.includes("413 request too large for model") && message.includes("tokens per minute (tpm)") && message.includes("limit ") && message.includes("requested ");
+}
+function wrapGroqOversizedRequestRecovery(streamFn, enabled) {
+	const underlying = streamFn ?? streamSimple;
+	if (!enabled) return underlying;
+	const withoutTools = (model, context, options) => {
+		const originalOnPayload = options?.onPayload;
+		return underlying(model, context, {
+			...options,
+			async onPayload(payload, payloadModel) {
+				const replacement = await originalOnPayload?.(payload, payloadModel);
+				const finalPayload = replacement && typeof replacement === "object" ? replacement : payload;
+				const record = finalPayload;
+				delete record.tools;
+				delete record.tool_choice;
+				delete record.parallel_tool_calls;
+				delete record.parallelToolCalls;
+				delete record.max_tokens;
+				delete record.max_completion_tokens;
+				record.max_completion_tokens = GROQ_FALLBACK_MAX_TOKENS;
+				return finalPayload;
+			}
+		});
+	};
+	return (model, context, options) => {
+		if (options?.maxTokens !== void 0) return underlying(model, context, options);
+		const initial = underlying(model, context, options);
+		const output = createAssistantMessageEventStream();
+		const writable = output;
+		(async () => {
+			try {
+				const resolvedInitial = await Promise.resolve(initial);
+				let forwarded = false;
+				let retryWithoutTools = false;
+				for await (const event of resolvedInitial) {
+					if (!forwarded && !options?.signal?.aborted && isGroqTpmRequestTooLargeEvent(event)) {
+						retryWithoutTools = true;
+						break;
+					}
+					writable.push(event);
+					forwarded = true;
+				}
+				if (retryWithoutTools) {
+					const fallback = await Promise.resolve(withoutTools(model, context, options));
+					for await (const event of fallback) writable.push(event);
+				}
+			} catch (error) {
+				writable.push({
+					type: "error",
+					reason: "error",
+					error: {
+						role: "assistant",
+						content: [],
+						api: model.api,
+						provider: model.provider,
+						model: model.id,
+						usage: {
+							input: 0,
+							output: 0,
+							cacheRead: 0,
+							cacheWrite: 0,
+							totalTokens: 0,
+							cost: {
+								input: 0,
+								output: 0,
+								cacheRead: 0,
+								cacheWrite: 0,
+								total: 0
+							}
+						},
+						stopReason: "error",
+						errorMessage: error instanceof Error ? error.message : String(error),
+						timestamp: Date.now()
+					}
+				});
+			} finally {
+				writable.end();
+			}
+		})();
+		return output;
+	};
+}
 var groq_default = definePluginEntry({
 	id: "groq",
 	name: "Groq Provider",
@@ -31,7 +277,20 @@ var groq_default = definePluginEntry({
 					groupLabel: "Groq",
 					groupHint: "Fast OpenAI-compatible inference"
 				}
-			})]
+			})],
+			catalog: {
+				order: "simple",
+				run: (ctx) => buildOpenAICompatibleProviderCatalog({
+					ctx,
+					providerId: "groq",
+					buildProvider: buildGroqCatalogProvider
+				})
+			},
+			staticCatalog: {
+				order: "simple",
+				run: async () => ({ provider: buildGroqCatalogProvider() })
+			},
+			wrapStreamFn: (ctx) => wrapGroqOversizedRequestRecovery(ctx.streamFn, ctx.modelId === GROQ_DEFAULT_MODEL_ID && !hasExplicitMaxTokens(ctx.extraParams) && !hasExplicitMaxTokens(ctx.model?.params) && ctx.model?.maxTokensSource === "discovered")
 		});
 		api.registerMediaUnderstandingProvider(groqMediaUnderstandingProvider);
 	}

package/npm-shrinkwrap.json

@@ -1,12 +1,12 @@
 {
   "name": "@openclaw/groq-provider",
-  "version": "2026.7.2-beta.2",
+  "version": "2026.7.2-beta.4",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "@openclaw/groq-provider",
-      "version": "2026.7.2-beta.2"
+      "version": "2026.7.2-beta.4"
     }
   }
 }

package/openclaw.plugin.json

@@ -182,7 +182,7 @@
       }
     },
     "discovery": {
-      "groq": "static"
+      "groq": "refreshable"
     }
   },
   "contracts": {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@openclaw/groq-provider",
-  "version": "2026.7.2-beta.2",
+  "version": "2026.7.2-beta.4",
   "description": "OpenClaw Groq media-understanding provider.",
   "repository": {
     "type": "git",
@@ -18,10 +18,10 @@
       "minHostVersion": ">=2026.6.8"
     },
     "compat": {
-      "pluginApi": ">=2026.7.2-beta.2"
+      "pluginApi": ">=2026.7.2-beta.4"
     },
     "build": {
-      "openclawVersion": "2026.7.2-beta.2",
+      "openclawVersion": "2026.7.2-beta.4",
       "bundledDist": false
     },
     "release": {
@@ -39,7 +39,7 @@
     "README.md"
   ],
   "peerDependencies": {
-    "openclaw": ">=2026.7.2-beta.2"
+    "openclaw": ">=2026.7.2-beta.4"
   },
   "peerDependenciesMeta": {
     "openclaw": {