@browserbasehq/browse-cli 0.4.0 → 0.4.1-alpha-3917df4

package/dist/index.js

@@ -100088,6 +100088,18 @@ Return the element that matches the instruction if it exists. Otherwise, return
     content: [content, buildUserInstructionsString(userProvidedInstructions)].filter(Boolean).join("\n\n")
   };
 }
+function buildActVariablesPrompt(variables) {
+  if (!variables || Object.keys(variables).length === 0) {
+    return "";
+  }
+  const variableNames = Object.keys(variables).map((key) => `%${key}%`).join(", ");
+  return ` The user has provided the following variables to be used in the action: ${variableNames} 
+
+    Note that these are the variable names/keys, and not the actual variable values. 
+
+    To use the variables in the action, you must respond with the variable name inside the 'arguments' array. The variable name must be wrapped in percentage signs (eg, %variableNameHere%) so that it can be replaced with the actual variable value before the action is taken. 
+`;
+}
 function buildActPrompt(action, supportedActions, variables) {
   let instruction = `Find the most relevant element to perform an action on given the following action: ${action}.  
   IF AND ONLY IF the action EXPLICITLY includes the word 'dropdown' and implies choosing/selecting an option from a dropdown, ignore the 'General Instructions' section, and follow the 'Dropdown Specific Instructions' section carefully.
@@ -100114,11 +100126,7 @@ function buildActPrompt(action, supportedActions, variables) {
         - choose the 'click' method
         - set twoStep to true.
   `;
-  if (variables && Object.keys(variables).length > 0) {
-    const variableNames = Object.keys(variables).map((key) => `%${key}%`).join(", ");
-    const variablesPrompt = `The following variables are available to use in the action: ${variableNames}. Fill the argument variables with the variable name.`;
-    instruction += ` ${variablesPrompt}`;
-  }
+  instruction += buildActVariablesPrompt(variables);
   return instruction;
 }
 function buildStepTwoPrompt(originalUserAction, previousAction, supportedActions, variables) {
@@ -100136,11 +100144,7 @@ function buildStepTwoPrompt(originalUserAction, previousAction, supportedActions
   If the user is asking to scroll to the next chunk/previous chunk, choose the nextChunk/prevChunk method. No arguments are required here.
   If the action implies a key press, e.g., 'press enter', 'press a', 'press space', etc., always choose the press method with the appropriate key as argument \u2014 e.g. 'a', 'Enter', 'Space'. Do not choose a click action on an on-screen keyboard. Capitalize the first character like 'Enter', 'Tab', 'Escape' only for special keys. 
   `;
-  if (variables && Object.keys(variables).length > 0) {
-    const variableNames = Object.keys(variables).map((key) => `%${key}%`).join(", ");
-    const variablesPrompt = `The following variables are available to use in the action: ${variableNames}. Fill the argument variables with the variable name.`;
-    instruction += ` ${variablesPrompt}`;
-  }
+  instruction += buildActVariablesPrompt(variables);
   return instruction;
 }
 function buildGoogleCUASystemPrompt() {
@@ -100448,7 +100452,7 @@ async function observe({ instruction, domElements, llmClient, userProvidedInstru
 async function act({ instruction, domElements, llmClient, userProvidedInstructions, logger, logInferenceToFile = false }) {
   const isGPT5 = llmClient.modelName.includes("gpt-5");
   const actSchema = external_exports.object({
-    elementId: external_exports.string().regex(/^\d+-\d+$/).describe("the ID string associated with the element. Never include surrounding square brackets. This field must follow the format of 'number-number'."),
+    elementId: external_exports.string().regex(/^\d+-\d+$/).describe("the ID string associated with the element. Never include surrounding square brackets. This field must follow the format of 'number-number'. for example, '0-76' or '16-21'"),
     description: external_exports.string().describe("a description of the accessible element and its purpose"),
     method: external_exports.enum(
       // Use Object.values() for Zod v3 compatibility - z.enum() in v3 doesn't accept TypeScript enums directly
@@ -119787,10 +119791,13 @@ var AISdkClient2 = class extends LLMClient {
   type = "aisdk";
   model;
   logger;
-  constructor({ model, logger }) {
+  constructor({ model, logger, clientOptions }) {
     super(model.modelId);
     this.model = model;
     this.logger = logger;
+    if (clientOptions) {
+      this.clientOptions = clientOptions;
+    }
   }
   getLanguageModel() {
     return this.model;
@@ -119868,9 +119875,11 @@ var AISdkClient2 = class extends LLMClient {
     let objectResponse;
     const isGPT5 = this.model.modelId.includes("gpt-5");
     const isCodex = this.model.modelId.includes("codex");
-    const usesLowReasoningEffort = (this.model.modelId.includes("gpt-5.1") || this.model.modelId.includes("gpt-5.2")) && !isCodex;
     const isKimi = this.model.modelId.includes("kimi");
     const temperature = isKimi ? 1 : options.temperature;
+    const isGPT5SubModel = this.model.modelId.includes("gpt-5.") && !isCodex;
+    const userReasoningEffort = this.clientOptions?.reasoningEffort;
+    const resolvedReasoningEffort = userReasoningEffort ?? (isGPT5SubModel ? "none" : void 0);
     const PROMPT_JSON_FALLBACK_PATTERNS = ["deepseek", "kimi", "glm"];
     const needsPromptJsonFallback = PROMPT_JSON_FALLBACK_PATTERNS.some((p2) => this.model.modelId.includes(p2));
     if (options.response_model) {
@@ -119899,11 +119908,10 @@ You must respond in JSON format. respond WITH JSON. Do not include any other tex
           messages: formattedMessages,
           schema: options.response_model.schema,
           temperature,
-          providerOptions: isGPT5 ? {
+          providerOptions: resolvedReasoningEffort ? {
             openai: {
-              textVerbosity: isCodex ? "medium" : "low",
-              // codex models only support 'medium'
-              reasoningEffort: isCodex ? "medium" : usesLowReasoningEffort ? "low" : "minimal"
+              ...isGPT5 ? { textVerbosity: isCodex ? "medium" : "low" } : {},
+              reasoningEffort: resolvedReasoningEffort
             }
           } : void 0
         });
@@ -152792,27 +152800,34 @@ var modelToProviderMap = {
   "gemini-2.5-flash-preview-04-17": "google",
   "gemini-2.5-pro-preview-03-25": "google"
 };
-function getAISDKLanguageModel(subProvider, subModelName, clientOptions) {
+function getAISDKLanguageModel(subProvider, subModelName, clientOptions, middleware) {
   const hasValidOptions = clientOptions && Object.values(clientOptions).some((v2) => v2 !== void 0 && v2 !== null);
+  let model;
   if (hasValidOptions) {
     const creator = AISDKProvidersWithAPIKey[subProvider];
     if (!creator) {
       throw new UnsupportedAISDKModelProviderError(subProvider, Object.keys(AISDKProvidersWithAPIKey));
     }
     const provider = creator(clientOptions);
-    return provider(subModelName);
+    model = provider(subModelName);
   } else {
     const provider = AISDKProviders[subProvider];
     if (!provider) {
       throw new UnsupportedAISDKModelProviderError(subProvider, Object.keys(AISDKProviders));
     }
-    return provider(subModelName);
+    model = provider(subModelName);
   }
+  if (middleware) {
+    return wrapLanguageModel({ model, middleware });
+  }
+  return model;
 }
 var LLMProvider = class {
   logger;
-  constructor(logger) {
+  middleware;
+  constructor(logger, middleware) {
     this.logger = logger;
+    this.middleware = middleware;
   }
   getClient(modelName, clientOptions, options) {
     if (modelName.includes("/")) {
@@ -152822,10 +152837,12 @@ var LLMProvider = class {
       if (subProvider === "vertex" && !options?.disableAPI && !options?.experimental) {
         throw new ExperimentalNotConfiguredError("Vertex provider");
       }
-      const languageModel = getAISDKLanguageModel(subProvider, subModelName, clientOptions);
+      const effectiveMiddleware = options?.middleware ?? this.middleware;
+      const languageModel = getAISDKLanguageModel(subProvider, subModelName, clientOptions, effectiveMiddleware);
       return new AISdkClient2({
         model: languageModel,
-        logger: this.logger
+        logger: this.logger,
+        clientOptions
       });
     }
     const provider = modelToProviderMap[modelName];
@@ -160612,7 +160629,9 @@ function hasInjectableDOM(url2) {
   return false;
 }
 function isNonWebTarget(info) {
-  return info.type !== "page" && info.type !== "iframe" || !hasInjectableDOM(info.url);
+  if (info.type === "page")
+    return false;
+  return info.type !== "iframe" || !hasInjectableDOM(info.url);
 }
 function isTopLevelPage(info) {
   const ti = info;
@@ -162828,13 +162847,14 @@ function resolveModelConfiguration(model) {
     return { modelName: model };
   }
   if (model && typeof model === "object") {
-    const { modelName, ...clientOptions } = model;
+    const { modelName, middleware, ...clientOptions } = model;
     if (!modelName) {
       throw new StagehandInvalidArgumentError("model.modelName is required when providing client options.");
     }
     return {
       modelName,
-      clientOptions
+      clientOptions,
+      middleware
     };
   }
   return { modelName: DEFAULT_MODEL_NAME };
@@ -163009,11 +163029,11 @@ var V3 = (() => {
         }
       } catch {
       }
-      const { modelName, clientOptions } = resolveModelConfiguration(opts.model);
+      const { modelName, clientOptions, middleware } = resolveModelConfiguration(opts.model);
       this.modelName = modelName;
       this.experimental = opts.experimental ?? false;
       this.logInferenceToFile = opts.logInferenceToFile ?? false;
-      this.llmProvider = new LLMProvider(this.logger);
+      this.llmProvider = new LLMProvider(this.logger, middleware);
       this.domSettleTimeoutMs = opts.domSettleTimeout;
       this.disableAPI = opts.disableAPI ?? false;
       const baseClientOptions = clientOptions ? { ...clientOptions } : {};
@@ -163095,14 +163115,16 @@ var V3 = (() => {
       }
       let modelName;
       let clientOptions;
+      let perCallMiddleware;
       if (typeof model === "string") {
         modelName = model;
       } else {
-        const { modelName: overrideModelName, ...rest } = model;
+        const { modelName: overrideModelName, middleware, ...rest } = model;
         modelName = overrideModelName;
         clientOptions = rest;
+        perCallMiddleware = middleware;
       }
-      if (modelName === this.modelName && (!clientOptions || Object.keys(clientOptions).length === 0)) {
+      if (modelName === this.modelName && !perCallMiddleware && (!clientOptions || Object.keys(clientOptions).length === 0)) {
         return this.llmClient;
       }
       const overrideProvider = String(modelName).split("/")[0];
@@ -163118,6 +163140,13 @@ var V3 = (() => {
           mergedOptions.apiKey = apiKey;
         }
       }
+      if (perCallMiddleware) {
+        return this.llmProvider.getClient(modelName, mergedOptions, {
+          experimental: this.experimental,
+          disableAPI: this.disableAPI,
+          middleware: perCallMiddleware
+        });
+      }
       const cacheKey = JSON.stringify({
         modelName,
         clientOptions: mergedOptions
@@ -163126,7 +163155,10 @@ var V3 = (() => {
       if (cached2) {
         return cached2;
       }
-      const client = this.llmProvider.getClient(modelName, mergedOptions, { experimental: this.experimental, disableAPI: this.disableAPI });
+      const client = this.llmProvider.getClient(modelName, mergedOptions, {
+        experimental: this.experimental,
+        disableAPI: this.disableAPI
+      });
       this.overrideLlmClients.set(cacheKey, client);
       return client;
     }
@@ -164620,7 +164652,7 @@ var import_child_process4 = require("child_process");
 var readline = __toESM(require("readline"));
 
 // package.json
-var version3 = "0.4.0";
+var version3 = "0.4.1";
 
 // src/resolve-ws.ts
 init_cjs_shims();
@@ -164645,6 +164677,7 @@ async function resolveWsTarget(input) {
 }
 
 // src/index.ts
+var import_node_html_markdown = require("node-html-markdown");
 var program = new import_commander.Command();
 var SOCKET_DIR = os3.tmpdir();
 function getSocketPath(session) {
@@ -165104,7 +165137,7 @@ async function runDaemon(session, headless) {
           } : {},
           ...!connectSessionId ? {
             browserbaseSessionCreateParams: {
-              userMetadata: { "browse-cli": "true" },
+              userMetadata: { browse_cli: "true" },
               ...contextConfig ? {
                 browserSettings: {
                   context: contextConfig
@@ -165466,6 +165499,11 @@ async function executeCommand(context, command, args, stagehand) {
           return {
             checked: await page.deepLocator(resolveSelector(selector)).isChecked()
           };
+        case "markdown": {
+          const target = selector ? resolveSelector(selector) : "body";
+          const html = await page.deepLocator(target).innerHtml();
+          return { markdown: import_node_html_markdown.NodeHtmlMarkdown.translate(html) };
+        }
         default:
           throw new Error(`Unknown get type: ${what}`);
       }
@@ -166395,7 +166433,7 @@ program.command("highlight <selector>").description("Highlight element").option(
   }
 });
 program.command("get <what> [selector]").description(
-  "Get page info: url, title, text, html, value, box, visible, checked"
+  "Get page info: url, title, text, html, markdown, value, box, visible, checked"
 ).action(async (what, selector) => {
   const opts = program.opts();
   try {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@browserbasehq/browse-cli",
-  "version": "0.4.0",
+  "version": "0.4.1-alpha-3917df4",
   "description": "Browser automation CLI for AI agents, built on Stagehand",
   "type": "commonjs",
   "license": "MIT",
@@ -44,10 +44,11 @@
   "dependencies": {
     "commander": "^12.0.0",
     "dotenv": "^16.4.5",
+    "node-html-markdown": "^1.3.0",
     "pino": "^9.6.0",
     "pino-pretty": "^13.0.0",
     "ws": "^8.18.0",
-    "@browserbasehq/stagehand": "3.2.1-alpha-be6798aefdf44d2d2d275a4b9cf3d4473609515b"
+    "@browserbasehq/stagehand": "3.2.0"
   },
   "devDependencies": {
     "@types/node": "^20.11.30",