Update setup-hf-config.mjs

setup-hf-config.mjs

@@ -3,19 +3,27 @@
  * One-time setup for OpenClaw on Hugging Face Spaces.
  * Runs at container startup; writes or merges openclaw.json from env (Secrets/Variables):
  *
- * Custom Ollama provider (hf-ollama-qwen3-vl) — PRIMARY default model:
+ * Model priority (highest wins):
+ *   1. GLM-4.7-Flash via Z.ai (ZHIPU_API_KEY set)  — fast cloud, free, 203K ctx
+ *   2. Custom Ollama HF Space (OPENCLAW_OLLAMA_BASE_URL set)  — local CPU, slow
+ *   3. HF Inference fallback (OPENCLAW_HF_DEFAULT_MODEL)  — DeepSeek-R1 default
+ *
+ * Z.ai / GLM provider:
+ * - providers["z-ai"] registered with GLM model catalog
+ * - agents.defaults.model.primary = "z-ai/glm-4.7-flash" when ZHIPU_API_KEY is set
+ * - Configure via: ZHIPU_API_KEY
+ *
+ * Custom Ollama provider (hf-ollama-qwen3-vl) — used when GLM key not set:
  * - providers["hf-ollama-qwen3-vl"] registered with baseUrl + apiKey
- * - agents.defaults.model.primary = "hf-ollama-qwen3-vl/voytas26/openclaw-qwen3vl-8b-opt"
  * - Configure via:
  *     OPENCLAW_OLLAMA_BASE_URL  (default: https://ubix-Clawd.hf.space/v1)
  *     OPENCLAW_OLLAMA_API_KEY   (default: "ollama")
- * - Falls back to OPENCLAW_HF_DEFAULT_MODEL (or HF DeepSeek-R1) when OPENCLAW_OLLAMA_BASE_URL
- *   is explicitly set to empty string to disable.
  *
- * OpenRouter / Perplexity (search-augmented model):
- * - providers.openrouter.apiKey from OPENROUTER_API_KEY
- * - agents.defaults.model.search set to openrouter/perplexity/sonar
- *   (only written when OPENROUTER_API_KEY is present)
+ * OpenRouter free model catalog (15 models registered as fallbacks):
+ * - providers.openrouter registered under models.providers with full model list
+ * - agents.defaults.model.fallbacks populated with all free tier models
+ * - Switch model in chat with: /model openrouter/<provider>/<model-id>
+ * - Only written when OPENROUTER_API_KEY is present
  *
  * Gateway auth:
  * - gateway.auth: OPENCLAW_GATEWAY_TOKEN (token) or OPENCLAW_GATEWAY_PASSWORD (password)
@@ -52,23 +60,28 @@ function readGatewayToken() {
 
 // ── Env reads ──────────────────────────────────────────────────────────────────────────────────
 
+// Z.ai / GLM API key — when set, GLM-4.7-Flash becomes primary (fast cloud, free)
+const zhipuApiKey = process.env.ZHIPU_API_KEY?.trim();
+
 // Custom Ollama/OpenAI-compat provider hosted on HF Space
-// Set OPENCLAW_OLLAMA_BASE_URL="" to disable and fall back to HF Inference
+// Set OPENCLAW_OLLAMA_BASE_URL="" to disable
 const ollamaBaseUrl = process.env.OPENCLAW_OLLAMA_BASE_URL !== undefined
   ? process.env.OPENCLAW_OLLAMA_BASE_URL.trim()
   : "https://ubix-Clawd.hf.space/v1";
 const ollamaApiKey  = process.env.OPENCLAW_OLLAMA_API_KEY?.trim() || "ollama";
 const ollamaEnabled = ollamaBaseUrl.length > 0;
 
-// Fallback HF Inference model (used only when Ollama provider is disabled)
+// Fallback HF Inference model (used only when both GLM and Ollama are disabled)
 const hfFallbackModel =
   process.env.OPENCLAW_HF_DEFAULT_MODEL?.trim() ||
   "huggingface/deepseek-ai/DeepSeek-R1";
 
-// Resolved primary model
-const defaultModel = ollamaEnabled
-  ? "hf-ollama-qwen3-vl/voytas26/openclaw-qwen3vl-8b-opt"
-  : hfFallbackModel;
+// Resolved primary model — GLM wins if key is set (fastest + free cloud API)
+const defaultModel = zhipuApiKey
+  ? "z-ai/glm-4.7-flash"
+  : ollamaEnabled
+    ? "hf-ollama-qwen3-vl/voytas26/openclaw-qwen3vl-8b-opt"
+    : hfFallbackModel;
 
 const gatewayToken    = readGatewayToken();
 const gatewayPassword = process.env.OPENCLAW_GATEWAY_PASSWORD?.trim();
@@ -118,7 +131,71 @@ if (!config.agents) config.agents = {};
 if (!config.agents.defaults) config.agents.defaults = {};
 if (!config.agents.defaults.model) config.agents.defaults.model = {};
 
-// 1a. Custom Ollama/OpenAI-compat provider (hf-ollama-qwen3-vl)
+// 1a. Z.ai / GLM provider — primary when key is set
+if (zhipuApiKey) {
+  config.models.providers["z-ai"] = {
+    baseUrl: "https://open.bigmodel.cn/api/paas/v4",
+    apiKey:  zhipuApiKey,
+    api:     "openai-completions",
+    models: [
+      // ── Free models ───────────────────────────────────────────────────────────────────────
+      {
+        id:            "glm-4.7-flash",
+        name:          "GLM-4.7-Flash (free) — fast, 203K ctx, tool calls",
+        reasoning:     true,
+        input:         ["text"],
+        contextWindow: 203000,
+        maxTokens:     16384,
+        cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+      },
+      {
+        id:            "glm-4.5-air",
+        name:          "GLM-4.5-Air (free) — MoE, thinking + tools",
+        reasoning:     true,
+        input:         ["text"],
+        contextWindow: 131072,
+        maxTokens:     16384,
+        cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+      },
+      {
+        id:            "glm-4v-flash",
+        name:          "GLM-4V-Flash (free) — vision + tools",
+        reasoning:     false,
+        input:         ["text", "image"],
+        contextWindow: 8192,
+        maxTokens:     4096,
+        cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+      },
+      // ── Paid models (low cost) ────────────────────────────────────────────────────────────
+      {
+        id:            "glm-4.7",
+        name:          "GLM-4.7 (paid) — full model, strongest reasoning",
+        reasoning:     true,
+        input:         ["text"],
+        contextWindow: 203000,
+        maxTokens:     16384,
+        cost: { input: 0.000003, output: 0.000003, cacheRead: 0, cacheWrite: 0 },
+      },
+      {
+        id:            "glm-4.5",
+        name:          "GLM-4.5 (paid) — MoE flagship, long context",
+        reasoning:     true,
+        input:         ["text"],
+        contextWindow: 131072,
+        maxTokens:     16384,
+        cost: { input: 0.000002, output: 0.000002, cacheRead: 0, cacheWrite: 0 },
+      },
+    ],
+  };
+  console.log("[openclaw-hf-setup] z-ai provider registered -> primary: glm-4.7-flash");
+} else {
+  console.warn(
+    "[openclaw-hf-setup] ZHIPU_API_KEY not set. GLM models unavailable. " +
+    "Add ZHIPU_API_KEY from https://open.bigmodel.cn to use GLM-4.7-Flash for free."
+  );
+}
+
+// 1b. Custom Ollama/OpenAI-compat provider (hf-ollama-qwen3-vl)
 if (ollamaEnabled) {
   config.models.providers["hf-ollama-qwen3-vl"] = {
     baseUrl: ollamaBaseUrl,
@@ -149,43 +226,193 @@ if (ollamaEnabled) {
 // 1b. Set primary default model
 config.agents.defaults.model.primary = defaultModel;
 
-// ── 2. OpenRouter provider + Perplexity Sonar as fallback model ───────────────────────────────
-// Note: model.search is not a valid key. Perplexity is registered as a provider and
-// added to model.fallbacks so it can be selected manually via /model in chat.
+// ── 2. OpenRouter provider + free model catalog as fallbacks ──────────────────────────────────
 if (openrouterKey) {
   config.models.providers.openrouter = {
-    apiKey: openrouterKey,
-    api:    "openai-completions",
+    apiKey:  openrouterKey,
+    api:     "openai-completions",
     baseUrl: "https://openrouter.ai/api/v1",
     models: [
+      // ── Web Search ─────────────────────────────────────────────────────────────────────────
       {
         id:            "perplexity/sonar",
-        name:          "Perplexity Sonar (web search)",
+        name:          "Perplexity Sonar (web search - paid)",
         reasoning:     false,
         input:         ["text"],
         contextWindow: 127072,
         maxTokens:     8192,
         cost: { input: 0.000001, output: 0.000001, cacheRead: 0, cacheWrite: 0 },
       },
+      // ── General Purpose (FREE) ────────────────────────────────────────────────────────────
+      {
+        id:            "meta-llama/llama-3.3-70b-instruct:free",
+        name:          "Llama 3.3 70B (free) — best general purpose",
+        reasoning:     false,
+        input:         ["text"],
+        contextWindow: 128000,
+        maxTokens:     8192,
+        cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+      },
+      {
+        id:            "qwen/qwen3-next-80b-a3b-instruct:free",
+        name:          "Qwen3 Next 80B (free) — strong general, 262K ctx",
+        reasoning:     false,
+        input:         ["text"],
+        contextWindow: 262144,
+        maxTokens:     16384,
+        cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+      },
+      {
+        id:            "mistralai/mistral-small-3.1-24b-instruct:free",
+        name:          "Mistral Small 3.1 24B (free) — vision + tools",
+        reasoning:     false,
+        input:         ["text", "image"],
+        contextWindow: 128000,
+        maxTokens:     8192,
+        cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+      },
+      {
+        id:            "google/gemma-3-27b-it:free",
+        name:          "Gemma 3 27B (free) — vision + tools",
+        reasoning:     false,
+        input:         ["text", "image"],
+        contextWindow: 131072,
+        maxTokens:     8192,
+        cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+      },
+      // ── Coding (FREE) ─────────────────────────────────────────────────────────────────────
+      {
+        id:            "qwen/qwen3-coder:free",
+        name:          "Qwen3 Coder 480B (free) — best free coding, 262K ctx",
+        reasoning:     false,
+        input:         ["text"],
+        contextWindow: 262144,
+        maxTokens:     16384,
+        cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+      },
+      {
+        id:            "openai/gpt-oss-120b:free",
+        name:          "OpenAI OSS 120B (free) — strong coding + tools",
+        reasoning:     false,
+        input:         ["text"],
+        contextWindow: 131072,
+        maxTokens:     16384,
+        cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+      },
+      // ── Reasoning (FREE) ──────────────────────────────────────────────────────────────────
+      {
+        id:            "qwen/qwen3-235b-a22b-thinking-2507",
+        name:          "Qwen3 235B Thinking (free) — best free reasoning",
+        reasoning:     true,
+        input:         ["text"],
+        contextWindow: 131072,
+        maxTokens:     16384,
+        cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+      },
+      {
+        id:            "arcee-ai/trinity-large-preview:free",
+        name:          "Arcee Trinity Large (free) — reasoning + tools",
+        reasoning:     true,
+        input:         ["text"],
+        contextWindow: 131072,
+        maxTokens:     8192,
+        cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+      },
+      {
+        id:            "stepfun/step-3.5-flash:free",
+        name:          "StepFun Step 3.5 Flash (free) — fast reasoning, 256K ctx",
+        reasoning:     true,
+        input:         ["text"],
+        contextWindow: 256000,
+        maxTokens:     16384,
+        cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+      },
+      // ── Vision + Multimodal (FREE) ────────────────────────────────────────────────────────
+      {
+        id:            "qwen/qwen3-vl-30b-a3b-thinking",
+        name:          "Qwen3 VL 30B Thinking (free) — vision + reasoning",
+        reasoning:     true,
+        input:         ["text", "image"],
+        contextWindow: 131072,
+        maxTokens:     8192,
+        cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+      },
+      {
+        id:            "nvidia/nemotron-nano-12b-v2-vl:free",
+        name:          "NVIDIA Nemotron 12B VL (free) — vision + tools",
+        reasoning:     false,
+        input:         ["text", "image"],
+        contextWindow: 128000,
+        maxTokens:     8192,
+        cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+      },
+      // ── Fast / Lightweight (FREE) ─────────────────────────────────────────────────────────
+      {
+        id:            "nvidia/nemotron-3-nano-30b-a3b:free",
+        name:          "NVIDIA Nemotron Nano 30B (free) — fast MoE, agentic",
+        reasoning:     false,
+        input:         ["text"],
+        contextWindow: 256000,
+        maxTokens:     16384,
+        cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+      },
+      {
+        id:            "z-ai/glm-4.5-air:free",
+        name:          "GLM-4.5 Air (free) — MoE, thinking + tools",
+        reasoning:     true,
+        input:         ["text"],
+        contextWindow: 131072,
+        maxTokens:     8192,
+        cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+      },
+      // ── Auto Router (FREE random) ─────────────────────────────────────────────────────────
+      {
+        id:            "openrouter/free",
+        name:          "OpenRouter Free Router — random free model",
+        reasoning:     false,
+        input:         ["text", "image"],
+        contextWindow: 200000,
+        maxTokens:     8192,
+        cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+      },
     ],
   };
 
-  // Add Perplexity as a fallback (not primary) — use /model in chat to switch
-  if (!config.agents.defaults.model.fallbacks) {
-    config.agents.defaults.model.fallbacks = [];
-  }
-  if (!config.agents.defaults.model.fallbacks.includes("openrouter/perplexity/sonar")) {
-    config.agents.defaults.model.fallbacks.push("openrouter/perplexity/sonar");
-  }
+  // Ordered fallback list — switch with /model in chat
+  config.agents.defaults.model.fallbacks = [
+    // GLM (Z.ai free) — add first if key is set
+    ...(zhipuApiKey ? [
+      "z-ai/glm-4.5-air",
+      "z-ai/glm-4v-flash",
+    ] : []),
+    // Web search
+    "openrouter/perplexity/sonar",
+    // General purpose
+    "openrouter/meta-llama/llama-3.3-70b-instruct:free",
+    "openrouter/qwen/qwen3-next-80b-a3b-instruct:free",
+    "openrouter/mistralai/mistral-small-3.1-24b-instruct:free",
+    // Coding
+    "openrouter/qwen/qwen3-coder:free",
+    "openrouter/openai/gpt-oss-120b:free",
+    // Reasoning
+    "openrouter/qwen/qwen3-235b-a22b-thinking-2507",
+    "openrouter/stepfun/step-3.5-flash:free",
+    // Vision
+    "openrouter/qwen/qwen3-vl-30b-a3b-thinking",
+    "openrouter/mistralai/mistral-small-3.1-24b-instruct:free",
+    // Auto
+    "openrouter/openrouter/free",
+  ];
 
   console.log(
-    "[openclaw-hf-setup] OpenRouter configured -> perplexity/sonar added as fallback model"
+    `[openclaw-hf-setup] OpenRouter configured -> ${config.models.providers.openrouter.models.length} models registered, ` +
+    `${config.agents.defaults.model.fallbacks.length} fallbacks set`
   );
 } else {
   console.warn(
     "[openclaw-hf-setup] OPENROUTER_API_KEY not set. " +
-    "Perplexity Sonar will not be available. " +
-    "Add OPENROUTER_API_KEY in Space Secrets to enable it."
+    "OpenRouter free models will not be available. " +
+    "Add OPENROUTER_API_KEY in Space Secrets to enable them."
   );
 }
 
@@ -233,10 +460,11 @@ const parts = [
   `token_present=${useTokenAuth ? "1" : "0"}`,
   `password_present=${usePasswordAuth ? "1" : "0"}`,
   `auth=${authKind}`,
+  `glm_configured=${zhipuApiKey ? "1" : "0"}`,
   `ollama_provider=${ollamaEnabled ? "1" : "0"}`,
   `default_model=${defaultModel}`,
   `openrouter_configured=${openrouterKey ? "1" : "0"}`,
-  `search_model=${openrouterKey ? "openrouter/perplexity/sonar (fallback)" : "none"}`,
+  `openrouter_models=${openrouterKey ? (config.models?.providers?.openrouter?.models?.length ?? 0) : 0}`,
   `trustedProxies=${trustedProxies.length}`,
   `allowedOrigins=${allowedOrigins.length}`,
 ];