testdriverai 6.0.32 → 6.1.1-canary.613bfa3.0

package/agent/events.js

@@ -1,5 +1,4 @@
 const { EventEmitter2 } = require("eventemitter2");
-const { censorSensitiveDataDeep } = require("./lib/censorship");
 
 // Factory function to create a new emitter instance with censoring middleware
 const createEmitter = () => {
@@ -13,14 +12,6 @@ const createEmitter = () => {
     ignoreErrors: false,
   });
 
-  // Override emit to censor sensitive data before emitting
-  const originalEmit = emitter.emit.bind(emitter);
-  emitter.emit = function (event, ...args) {
-    // Censor all arguments passed to emit
-    const censoredArgs = args.map(censorSensitiveDataDeep);
-    return originalEmit(event, ...censoredArgs);
-  };
-
   return emitter;
 };
 
@@ -46,7 +37,7 @@ const events = {
   status: "status",
   log: {
     markdown: {
-      static: "log:markdown:static",
+      static: "log:markdown",
       start: "log:markdown:start",
       chunk: "log:markdown:chunk",
       end: "log:markdown:end",

package/agent/index.js

@@ -63,7 +63,11 @@ class TestDriverAgent extends EventEmitter2 {
     // Derive properties from cliArgs
     const flags = cliArgs.options || {};
     const firstArg = cliArgs.args && cliArgs.args[0];
+
+    // All commands (run, edit, generate) use the same pattern:
+    // first argument is the main file to work with
     this.thisFile = firstArg || this.config.TD_DEFAULT_TEST_FILE;
+
     this.resultFile = flags.resultFile || null;
     this.newSandbox = flags.newSandbox || false;
     this.healMode = flags.healMode || flags.heal || false;
@@ -427,6 +431,7 @@ class TestDriverAgent extends EventEmitter2 {
 
     // Log current execution position for debugging
     if (this.sourceMapper.currentFileSourceMap) {
+      this.emitter.emit(events.log.log, "");
       this.emitter.emit(
         events.log.log,
         theme.dim(`${this.sourceMapper.getCurrentPositionDescription()}`),
@@ -886,30 +891,30 @@ commands:
   // based on the current state of the system (primarily the current screenshot)
   // it will generate files that contain only "prompts"
   // @todo revit the generate command
-  async generate(type, count, baseYaml, skipYaml = false) {
-    this.emitter.emit(events.log.debug, "generate called, %s", type);
+  async generate(count = 1) {
+    this.emitter.emit(events.log.debug, `generate called with count: ${count}`);
 
-    this.emitter.emit(events.log.narration, theme.dim("thinking..."), true);
+    await this.runLifecycle("prerun");
 
-    if (baseYaml && !skipYaml) {
-      await this.runLifecycle("prerun");
-      await this.run(baseYaml, false, false);
-      await this.runLifecycle("postrun");
-    }
+    this.emitter.emit(events.log.narration, theme.dim("thinking..."), true);
 
     let image = await this.system.captureScreenBase64();
 
     const streamId = `generate-${Date.now()}`;
     this.emitter.emit(events.log.markdown.start, streamId);
 
+    let mouse = await this.system.getMousePosition();
+    let activeWindow = await this.system.activeWin();
+
     let message = await this.sdk.req(
       "generate",
       {
-        type,
+        prompt: "make sure to do a spellcheck",
         image,
-        mousePosition: await this.system.getMousePosition(),
-        activeWindow: await this.system.activeWin(),
+        mousePosition: mouse,
+        activeWindow: activeWindow,
         count,
+        stream: false,
       },
       (chunk) => {
         if (chunk.type === "data") {
@@ -932,35 +937,36 @@ commands:
           .replace(/['"`]/g, "")
           .replace(/[^a-zA-Z0-9-]/g, "") // remove any non-alphanumeric chars except hyphens
           .toLowerCase() + ".yaml";
+
       let path1 = path.join(
         this.workingDir,
         "testdriver",
         "generate",
         fileName,
       );
-
       // create generate directory if it doesn't exist
-      if (!fs.existsSync(path.join(this.workingDir, "generate"))) {
-        fs.mkdirSync(path.join(this.workingDir, "generate"));
+      const generateDir = path.join(this.workingDir, "testdriver", "generate");
+      if (!fs.existsSync(generateDir)) {
+        fs.mkdirSync(generateDir);
+        console.log("Created generate directory:", generateDir);
+      } else {
+        console.log("Generate directory already exists:", generateDir);
       }
 
       let list = testPrompt.steps;
 
-      if (baseYaml && fs.existsSync(baseYaml)) {
-        list.unshift({
-          step: {
-            command: "run",
-            file: baseYaml,
-          },
-        });
-      }
       let contents = yaml.dump({
         version: packageJson.version,
         steps: list,
       });
+
+      this.emitter.emit(events.log.debug, `writing file ${path1} ${contents}`);
+
       fs.writeFileSync(path1, contents);
     }
 
+    await this.runLifecycle("postrun");
+
     this.exit(false);
   }
 
@@ -1511,6 +1517,8 @@ ${regression}
   }
 
   async embed(file, depth, pushToHistory) {
+    let inputFile = JSON.parse(JSON.stringify(file));
+
     this.analytics.track("embed", { file });
 
     this.emitter.emit(
@@ -1520,7 +1528,7 @@ ${regression}
 
     depth = depth + 1;
 
-    this.emitter.emit(events.log.log, `${file} (start)`);
+    this.emitter.emit(events.log.log, `${inputFile} (start)`);
 
     // Use the new helper method to resolve file paths relative to testdriver directory
     const currentFilePath = this.sourceMapper.currentFilePath || this.thisFile;
@@ -1573,7 +1581,7 @@ ${regression}
       this.sourceMapper.restoreContext(previousContext);
     }
 
-    this.emitter.emit(events.log.log, `${file} (end)`);
+    this.emitter.emit(events.log.log, `${inputFile} (end)`);
   }
 
   // Returns sandboxId to use (either from file if recent, or null)
@@ -1780,10 +1788,6 @@ ${regression}
       // Start the debugger server as early as possible to ensure event listeners are attached
       if (!debuggerStarted) {
         debuggerStarted = true; // Prevent multiple starts, especially when running test in parallel
-        this.emitter.emit(
-          events.log.narration,
-          theme.green(`Starting debugger server...`),
-        );
         debuggerProcess = await createDebuggerProcess(
           this.config,
           this.emitter,
@@ -1791,6 +1795,7 @@ ${regression}
       }
       this.debuggerUrl = debuggerProcess.url || null; // Store the debugger URL
       this.emitter.emit(events.log.log, `This is beta software!`);
+      this.emitter.emit(events.log.log, ``);
       this.emitter.emit(
         events.log.log,
         theme.yellow(`Join our Discord for help`),
@@ -1799,6 +1804,7 @@ ${regression}
         events.log.log,
         `https://discord.com/invite/cWDFW8DzPm`,
       );
+      this.emitter.emit(events.log.log, ``);
 
       // make testdriver directory if it doesn't exist
       let testdriverFolder = path.join(this.workingDir);
@@ -1812,7 +1818,10 @@ ${regression}
       }
 
       // if the directory for thisFile doesn't exist, create it
-      if (this.cliArgs.command !== "sandbox") {
+      if (
+        this.cliArgs.command !== "sandbox" &&
+        this.cliArgs.command !== "generate"
+      ) {
         const dir = path.dirname(this.thisFile);
         if (!fs.existsSync(dir)) {
           fs.mkdirSync(dir, { recursive: true });
@@ -1837,7 +1846,10 @@ ${regression}
         await this.sdk.auth();
       }
 
-      if (this.cliArgs.command !== "sandbox") {
+      if (
+        this.cliArgs.command !== "sandbox" &&
+        this.cliArgs.command !== "generate"
+      ) {
         this.emitter.emit(
           events.log.log,
           theme.dim(`Working on ${this.thisFile}`),
@@ -2034,6 +2046,20 @@ Please check your network connection, TD_API_KEY, or the service status.`,
     // Use the current file path from sourceMapper to find the lifecycle directory
     // If sourceMapper doesn't have a current file, use thisFile which should be the file being run
     let currentFilePath = this.sourceMapper.currentFilePath || this.thisFile;
+
+    this.emitter.emit(events.log.log, ``);
+    this.emitter.emit(events.log.log, "Running lifecycle: " + lifecycleName);
+
+    // If we still don't have a currentFilePath, fall back to the default testdriver directory
+    if (!currentFilePath) {
+      currentFilePath = path.join(
+        this.workingDir,
+        "testdriver",
+        "testdriver.yaml",
+      );
+      console.log("No currentFilePath found, using fallback:", currentFilePath);
+    }
+
     // Ensure we have an absolute path
     if (currentFilePath && !path.isAbsolute(currentFilePath)) {
       currentFilePath = path.resolve(this.workingDir, currentFilePath);
@@ -2070,6 +2096,9 @@ Please check your network connection, TD_API_KEY, or the service status.`,
         }
       }
     }
+
+    this.emitter.emit(events.log.log, lifecycleFile);
+
     if (lifecycleFile) {
       // Store current source mapping state before running lifecycle file
       const previousContext = this.sourceMapper.saveContext();
@@ -2139,7 +2168,7 @@ Please check your network connection, TD_API_KEY, or the service status.`,
     }
 
     // Move environment setup and special handling here
-    if (["edit", "run"].includes(commandName)) {
+    if (["edit", "run", "generate"].includes(commandName)) {
       await this.buildEnv(options);
     }
 

package/agent/interface.js

@@ -196,6 +196,42 @@ function createCommandDefinitions(agent) {
         console.log(`TestDriver.ai v${packageJson.version}`);
       },
     },
+
+    generate: {
+      description: "Generate test files based on current screen state",
+      args: {
+        file: Args.string({
+          description: "Base test file to run before generating (optional)",
+          required: false,
+        }),
+      },
+      flags: {
+        count: Flags.integer({
+          description: "Number of test files to generate",
+          default: 3,
+        }),
+        headless: Flags.boolean({
+          description: "Run in headless mode (no GUI)",
+          default: false,
+        }),
+        new: Flags.boolean({
+          description:
+            "Create a new sandbox instead of reconnecting to an existing one",
+          default: false,
+        }),
+        "sandbox-ami": Flags.string({
+          description: "Specify AMI ID for sandbox instance (e.g., ami-1234)",
+        }),
+        "sandbox-instance": Flags.string({
+          description: "Specify EC2 instance type for sandbox (e.g., i3.metal)",
+        }),
+      },
+      handler: async (args, flags) => {
+        // The file argument is already handled by thisFile in the agent constructor
+        // Just call generate with the count
+        await agent.generate(flags.count || 3);
+      },
+    },
   };
 }
 

package/agent/lib/censorship.js

@@ -38,18 +38,23 @@ const censorSensitiveData = (message) => {
 
 // Function to censor sensitive data in any value (recursive for objects/arrays)
 const censorSensitiveDataDeep = (value) => {
-  if (typeof value === "string") {
-    return censorSensitiveData(value);
-  } else if (Array.isArray(value)) {
-    return value.map(censorSensitiveDataDeep);
-  } else if (value && typeof value === "object") {
-    const result = {};
-    for (const [key, val] of Object.entries(value)) {
-      result[key] = censorSensitiveDataDeep(val);
+  try {
+    if (typeof value === "string") {
+      return censorSensitiveData(value);
+    } else if (Array.isArray(value)) {
+      return value.map(censorSensitiveDataDeep);
+    } else if (value && typeof value === "object") {
+      const result = {};
+      for (const [key, val] of Object.entries(value)) {
+        result[key] = censorSensitiveDataDeep(val);
+      }
+      return result;
     }
-    return result;
+    return value;
+  } catch {
+    // If we hit any error (like circular reference), just return a safe placeholder
+    return "[Object]";
   }
-  return value;
 };
 
 // Function to update interpolation variables (for runtime updates)

package/agent/lib/commander.js

@@ -80,21 +80,21 @@ commands:
       // this will actually interpret the command and execute it
       switch (object.command) {
         case "type":
-          emitter.emit(events.log.narration, `typing ${object.text}`);
           emitter.emit(events.log.log, generator.jsonToManual(object));
+          emitter.emit(events.log.narration, `typing ${object.text}`);
           response = await commands.type(object.text, object.delay);
           break;
         case "press-keys":
+          emitter.emit(events.log.log, generator.jsonToManual(object));
           emitter.emit(
             events.log.narration,
-            `pressing keys ${object.keys.join(",")}`,
+            `pressing keys: ${Array.isArray(object.keys) ? object.keys.join(", ") : object.keys}`,
           );
-          emitter.emit(events.log.log, generator.jsonToManual(object));
           response = await commands["press-keys"](object.keys);
           break;
         case "scroll":
-          emitter.emit(events.log.narration, `scrolling ${object.direction}`);
           emitter.emit(events.log.log, generator.jsonToManual(object));
+          emitter.emit(events.log.narration, `scrolling ${object.direction}`);
           response = await commands.scroll(
             object.direction,
             object.amount,
@@ -102,21 +102,21 @@ commands:
           );
           break;
         case "wait":
+          emitter.emit(events.log.log, generator.jsonToManual(object));
           emitter.emit(
             events.log.narration,
             `waiting ${object.timeout} seconds`,
           );
-          emitter.emit(events.log.log, generator.jsonToManual(object));
           response = await commands.wait(object.timeout);
           break;
         case "click":
-          emitter.emit(events.log.narration, `${object.action}`);
           emitter.emit(events.log.log, generator.jsonToManual(object));
+          emitter.emit(events.log.narration, `${object.action}`);
           response = await commands["click"](object.x, object.y, object.action);
           break;
         case "hover":
-          emitter.emit(events.log.narration, `moving mouse`);
           emitter.emit(events.log.log, generator.jsonToManual(object));
+          emitter.emit(events.log.narration, `moving mouse`);
           response = await commands["hover"](object.x, object.y);
           break;
         case "drag":
@@ -124,11 +124,11 @@ commands:
           response = await commands["drag"](object.x, object.y);
           break;
         case "hover-text":
+          emitter.emit(events.log.log, generator.jsonToManual(object));
           emitter.emit(
             events.log.narration,
             `searching for ${object.description}`,
           );
-          emitter.emit(events.log.log, generator.jsonToManual(object));
           response = await commands["hover-text"](
             object.text,
             object.description,
@@ -138,11 +138,11 @@ commands:
           );
           break;
         case "hover-image":
+          emitter.emit(events.log.log, generator.jsonToManual(object));
           emitter.emit(
             events.log.narration,
             `searching for image of ${object.description}`,
           );
-          emitter.emit(events.log.log, generator.jsonToManual(object));
           response = await commands["hover-image"](
             object.description,
             object.action,
@@ -157,19 +157,19 @@ commands:
           response = await commands["match-image"](object.path, object.action);
           break;
         case "wait-for-image":
+          emitter.emit(events.log.log, generator.jsonToManual(object));
           emitter.emit(
             events.log.narration,
             `waiting for ${object.description}`,
           );
-          emitter.emit(events.log.log, generator.jsonToManual(object));
           response = await commands["wait-for-image"](
             object.description,
             object.timeout,
           );
           break;
         case "wait-for-text":
-          emitter.emit(events.log.narration, `waiting for ${object.text}`);
           emitter.emit(events.log.log, generator.jsonToManual(object));
+          emitter.emit(events.log.narration, `waiting for ${object.text}`);
           copy.text = "*****";
           response = await commands["wait-for-text"](
             object.text,
@@ -178,8 +178,8 @@ commands:
           );
           break;
         case "scroll-until-text":
-          emitter.emit(events.log.narration, `scrolling until ${object.text}`);
           emitter.emit(events.log.log, generator.jsonToManual(object));
+          emitter.emit(events.log.narration, `scrolling until ${object.text}`);
           copy.text = "*****";
           response = await commands["scroll-until-text"](
             object.text,
@@ -191,8 +191,8 @@ commands:
           break;
         case "scroll-until-image": {
           const needle = object.description || object.path;
-          emitter.emit(events.log.narration, `scrolling until ${needle}`);
           emitter.emit(events.log.log, generator.jsonToManual(object));
+          emitter.emit(events.log.narration, `scrolling until ${needle}`);
           response = await commands["scroll-until-image"](
             object.description,
             object.direction,
@@ -203,8 +203,8 @@ commands:
           break;
         }
         case "focus-application":
-          emitter.emit(events.log.narration, `focusing ${object.name}`);
           emitter.emit(events.log.log, generator.jsonToManual(object));
+          emitter.emit(events.log.narration, `focusing ${object.name}`);
           response = await commands["focus-application"](object.name);
           break;
         case "remember": {
@@ -215,12 +215,12 @@ commands:
           break;
         }
         case "assert":
-          emitter.emit(events.log.narration, `asserting ${object.expect}`);
           emitter.emit(events.log.log, generator.jsonToManual(object));
+          emitter.emit(events.log.narration, `asserting ${object.expect}`);
           response = await commands.assert(object.expect, object.async);
+
           break;
         case "exec":
-          emitter.emit(events.log.narration, `exec`);
           emitter.emit(
             events.log.log,
             generator.jsonToManual({

package/agent/lib/commands.js

@@ -176,7 +176,7 @@ const createCommands = (
     }
 
     const handleAssertResponse = (response) => {
-      emitter.emit(events.log.markdown.static, response);
+      emitter.emit(events.log.log, response);
 
       if (response.indexOf("The task passed") > -1) {
         return true;
@@ -727,14 +727,14 @@ const createCommands = (
             `Command failed with exit code ${result.out.returncode}: ${result.out.stderr}`,
           );
         } else {
-          if (!silent) {
-            emitter.emit(events.log.log, theme.dim(`Command stdout:`), true);
+          if (!silent && result.out?.stdout) {
+            emitter.emit(events.log.log, theme.dim(`stdout:`), true);
             emitter.emit(events.log.log, `${result.out.stdout}`, true);
+          }
 
-            if (result.out.stderr) {
-              emitter.emit(events.log.log, theme.dim(`Command stderr:`), true);
-              emitter.emit(events.log.log, `${result.out.stderr}`, true);
-            }
+          if (!silent && result.out.stderr) {
+            emitter.emit(events.log.log, theme.dim(`stderr:`), true);
+            emitter.emit(events.log.log, `${result.out.stderr}`, true);
           }
 
           return result.out?.stdout?.trim();

package/agent/lib/debugger-server.js

@@ -65,11 +65,6 @@ function createDebuggerServer(config = {}) {
 
       ws.on("close", () => {
         clients.delete(ws);
-
-        // If no clients connected, we can optionally shut down
-        if (clients.size === 0) {
-          console.log("No clients connected, keeping server alive");
-        }
       });
 
       ws.on("error", (error) => {

package/agent/lib/generator.js

@@ -1,6 +1,6 @@
 // parses markdown content to find code blocks, and then extracts yaml from those code blocks
 const yaml = require("js-yaml");
-const package = require("../../package.json");
+const pkg = require("../../package.json");
 const session = require("./session");
 const theme = require("./theme");
 // do the actual parsing
@@ -56,7 +56,7 @@ const jsonToManual = function (json, colors = true) {
 const dumpToYML = async function (inputArray, sessionInstance = null) {
   // use yml dump to convert json to yml
   let yml = await yaml.dump({
-    version: package.version,
+    version: pkg.version,
     session: sessionInstance ? sessionInstance.get() : session.get(),
     steps: inputArray,
   });

package/agent/lib/sdk.js

@@ -95,7 +95,7 @@ const createSDK = (emitter, config, sessionInstance) => {
         return token;
       } catch (error) {
         outputError(error);
-        return;
+        throw error; // Re-throw the error so calling code can handle it properly
       }
     }
   };
@@ -194,6 +194,7 @@ const createSDK = (emitter, config, sessionInstance) => {
       return value;
     } catch (error) {
       outputError(error);
+      throw error; // Re-throw the error so calling code can handle it properly
     }
   };
 

package/agent/lib/source-mapper.js

@@ -263,7 +263,7 @@ class SourceMapper {
     let description = `${fileName}:${(sourcePosition.step.startLine || 0) + 1}`;
 
     if (sourcePosition.command) {
-      description += `:${(sourcePosition.command.startLine || 0) + 1} (${sourcePosition.command.command || "unknown command"})`;
+      description += `:${(sourcePosition.command.startLine || 0) + 1}`;
     } else {
       description += ` (step ${sourcePosition.step.stepIndex + 1})`;
     }

package/docs/apps/tauri-apps.mdx

@@ -0,0 +1,413 @@
+---
+title: "Tauri Apps"
+sidebarTitle: "Tauri Apps"
+description: "Testing Tauri Apps with TestDriver"
+icon: "https://tauri.app/favicon.svg"
+---
+
+> [Tauri](https://tauri.app/) is a framework for building tiny, fast binaries for all major desktop and mobile platforms. Developers can integrate any frontend framework that compiles to HTML, JavaScript, and CSS for building their user experience while leveraging languages such as Rust, Swift, and Kotlin for backend logic when needed.
+>
+> <br />– [https://tauri.app/start](https://tauri.app/start/)
+
+## Testing Tauri apps with TestDriver
+
+In this guide, we'll leverage [Playwright](https://playwright.dev/) and the
+[TestDriver Playwright SDK](/getting-started/playwright) to convert the [Tauri Quick Start](https://tauri.app/start/create-project/)
+to TestDriver's selectorless, Vision AI.
+
+<Callout icon="code">
+  View Source:
+  [https://github.com/testdriverai/demo-tauri-app](https://github.com/testdriverai/demo-tauri-app)
+</Callout>
+
+### Requirements
+
+To start testing your Tauri app with TestDriver, you need the following:
+
+<AccordionGroup>
+  <Accordion title="Create a TestDriver account">
+    <Steps>
+      <Step title="Create a TestDriver Account">
+        You will need a [TestDriver Pro](https://app.testdriver.ai/team) account (\$20/month) to get an API key.
+        <Card title="Sign Up for TestDriver" icon="user-plus" horizontal href="https://app.testdriver.ai/team" />
+      </Step>
+      <Step title="Set up your environment">
+        Copy your API key from [the TestDriver dashboard](https://app.testdriver.ai/team), and set it as an environment variable.
+
+        <Tabs>
+          <Tab title="macOS / Linux">
+            Export an environment variable on macOS or Linux systems:
+
+            ```bash
+            export TD_API_KEY="your_api_key_here"
+            ```
+          </Tab>
+          <Tab title="Windows">
+            Export an environment variable in PowerShell:
+
+            ```powershell
+            setx TD_API_KEY "your_api_key_here"
+            ```
+          </Tab>
+        </Tabs>
+      </Step>
+
+    </Steps>
+
+  </Accordion>
+
+  <Accordion title="Create a Tauri project">
+
+    <Note>
+      Follow Tauri's [Create a Project](https://tauri.app/start/create-project/)
+      guide.
+    </Note>
+
+  </Accordion>
+
+    <Accordion title="Create a Playwright project">
+    <Info>
+      This is a condensed version of [Playwright's Installation Instructions](https://playwright.dev/docs/intro).
+
+      **If you're new to Playwright, you should follow their guide first.**
+    </Info>
+
+    In your Tauri project, run:
+
+    <Tabs>
+      <Tab title="npm">
+        ```bash
+        npm init playwright@latest
+        ```
+      </Tab>
+      <Tab title="yarn">
+        ```bash
+        yarn create playwright
+        ```
+      </Tab>
+      <Tab title="pnpm">
+        ```bash
+        pnpm create playwright
+        ```
+      </Tab>
+    </Tabs>
+    Select the following options when prompted:
+
+    ```console
+    ✔ Do you want to use TypeScript or JavaScript?
+    > TypeScript
+    ✔ Where to put your end-to-end tests?
+    > tests
+    ✔ Add a GitHub Actions workflow? (y/N)
+    > N
+    ✔ Install Playwright browsers (can be done manually via 'npx playwright install')? (Y/n)
+    > Y
+    ```
+
+  </Accordion>
+
+  <Accordion title="Install the TestDriver Playwright SDK">
+    `@testdriver.ai/playwright` is an AI-powered extension of `@playwright/test`.
+
+    <Tabs>
+      <Tab title="npm">
+        ```bash
+        npm install @testdriver.ai/playwright
+        ```
+      </Tab>
+      <Tab title="yarn">
+        ```bash
+        yarn add @testdriver.ai/playwright
+        ```
+      </Tab>
+      <Tab title="pnpm">
+        ```bash
+        pnpm add @testdriver.ai/playwright
+        ```
+      </Tab>
+    </Tabs>
+
+  </Accordion>
+
+</AccordionGroup>
+
+## Testing the Tauri Web App
+
+### Setup
+
+First, we need to modify the default Playwright configuration and our Tauri project to work together:
+
+<Steps>
+  <Step title="Configure Playwright to start the Tauri frontend">
+
+    In the `playwright.config.ts` file, we'll configure the [`webServer`](https://playwright.dev/docs/test-webserver)
+    to start the Tauri frontend for Playwright to test against:
+
+    ```typescript playwright.config.ts
+      /* Run your local dev server before starting the tests */
+      // [!code ++:8]
+      webServer: {
+        command: "npm run dev",
+        url: "http://localhost:1420",
+        reuseExistingServer: !process.env.CI,
+        env: {
+          VITE_PLAYWRIGHT: "true",
+        },
+      },
+    });
+    ```
+
+  </Step>
+
+  <Step title="Mock Tauri APIs">
+
+    Since we're testing the Tauri frontend, we need to [mock IPC Requests](https://tauri.app/develop/tests/mocking/)
+    to simulate `invoke` calls to the Rust backend:
+
+    ```html src/index.html
+    <body>
+      <div id="root"></div>
+      <!-- [!code ++:7] -->
+      <script type="module">
+        // This is set in playwright.config.ts to allow our tests to mock Tauri IPC `invoke` calls
+        if (import.meta.env.VITE_PLAYWRIGHT) {
+          const { mockIPC } = await import("@tauri-apps/api/mocks");
+          window.mockIPC = mockIPC;
+        }
+      </script>
+      <script type="module" src="/src/main.tsx"></script>
+    </body>
+    ```
+
+    We only need to do this once, as we'll be accessing `window.mockIPC` in our tests.
+
+  </Step>
+
+  <Step title="Create a new test file">
+    Create a new file (e.g. `tests/testdriver.spec.ts`) with:
+
+    ```typescript tests/testdriver.spec.ts
+    import type { mockIPC } from "@tauri-apps/api/mocks";
+    import { expect, test } from "@playwright/test";
+
+    test.beforeEach(async ({ page }) => {
+      await page.goto("http://localhost:1420");
+    });
+
+    test("should have title", async ({ page }) => {
+      await expect(page).toHaveTitle("Tauri + React + TypeScript");
+    });
+    ```
+
+  </Step>
+
+  <Step title="Run Playwright in UI Mode">
+
+    Now we're ready to run Playwright and start working on our tests:
+
+    <Tabs>
+      <Tab title="npm">
+
+        ```bash
+        npx playwright test --ui
+        ```
+
+      </Tab>
+      <Tab title="yarn">
+
+        ```bash
+        yarn playwright test --ui
+        ```
+
+      </Tab>
+      <Tab title="pnpm">
+
+        ```bash
+        pnpm exec playwright test --ui
+        ```
+
+      </Tab>
+    </Tabs>
+
+    ![Playwright UI Mode](https://playwright.dev/assets/ideal-img/ui-mode.4e54d6b.3598.png)
+
+    Click the <Icon icon="play" /> button to successfully run the tests in the UI.
+
+    <Tip>
+      Click the <Icon icon="eye" /> button to automatically re-run tests on save.
+    </Tip>
+
+  </Step>
+
+</Steps>
+
+### Usage
+
+#### Import the TestDriver Playwright SDK
+
+By changing **1 line**, we can add TestDriver's AI capabilities to Playwright:
+
+```typescript tests/testdriver.spec.ts
+import type { mockIPC } from "@tauri-apps/api/mocks";
+// [!code --]
+import { expect, test } from "@playwright/test";
+// [!code ++]
+import { expect, test } from "@testdriver.ai/playwright";
+
+// For type-safety of the mockIPC function
+declare global {
+  interface Window {
+    mockIPC: typeof mockIPC;
+  }
+}
+
+test.beforeEach(async ({ page }) => {
+  await page.goto("http://localhost:1420");
+});
+
+test("should have title", async ({ page }) => {
+  await expect(page).toHaveTitle("Tauri + React + TypeScript");
+});
+```
+
+Notice how we're able to continue using Playwright's API (`toHaveTitle`)
+with `@testdriver.ai/playwright`.
+
+The test continues to pass as before, so now we can update our test to use natural language instead of selectors.
+
+#### Assertions with `expect.toMatchPrompt`
+
+With Playwright, we would normally use a `getByRole` selector to assert the heading text:
+
+```typescript tests/example.spec.ts
+test("should have heading", async ({ page }) => {
+  await expect(
+    page.getByRole("heading", { name: "Installation" }),
+  ).toBeVisible();
+});
+```
+
+With TestDriver, we can use natural language with `toMatchPrompt` instead:
+
+```typescript tests/testdriver.spec.ts
+test("should have heading", async ({ page }) => {
+  // [!code --:3]
+  await expect(
+    page.getByRole("heading", { name: "Installation" }),
+  ).toBeVisible();
+  // [!code ++:1]
+  await expect(page).toMatchPrompt("Heading says 'Welcome to Tauri + React'");
+});
+```
+
+#### Agentic tests with `test.agent`
+
+With TestDriver, we can skip the test implementation **entirely** and let AI perform the test for us:
+
+<Steps>
+  <Step title="Mock the `greet` call">
+
+    First, we need to [mock our `invoke` calls](https://tauri.app/develop/tests/mocking/#ipc-requests),
+    since we're testing the frontend behavior and not our Tauri backend:
+
+    ```typescript tests/testdriver.spec.ts
+    test.beforeEach(async ({ page }) => {
+      await page.goto("http://localhost:1420");
+      // [!code ++:12]
+      await page.evaluate(() => {
+        // https://tauri.app/develop/tests/mocking/#mocking-commands-for-invoke
+        window.mockIPC((cmd, args) => {
+          switch (cmd) {
+            case "greet":
+              args = args as { name: string };
+              return `Hello, ${args.name}! You've been greeted from Rust!`;
+            default:
+              throw new Error(`Unsupported command: ${cmd}`);
+          }
+        });
+      });
+    });
+    ```
+
+  </Step>
+
+  <Step title="Add an Agentic Test">
+
+    Next, wrap a _prompt_ in `test.agent` to perform the test:
+
+    ```typescript tests/testdriver.spec.ts
+    test.describe("should greet with name", () => {
+      test.agent(`
+        - Enter the name "Tauri"
+        - Click the "Greet" button
+        - You should see the text "Hello, Tauri! You've been greeted from Rust!"
+      `);
+    });
+    ```
+
+  </Step>
+</Steps>
+
+#### Continued Reading
+
+[Learn more about TestDriver's Playwright SDK](/getting-started/playwright)
+
+## Testing the Tauri Desktop App
+
+We can use TestDriver and natural language to test our Tauri desktop app:
+
+<Steps>
+  <Step title="Run the Desktop App">
+    <Tabs>
+      <Tab title="npm">
+        ```bash
+        npm run tauri dev
+        ```
+      </Tab>
+      <Tab title="yarn">
+        ```bash
+        yarn tauri dev
+        ```
+      </Tab>
+      <Tab title="pnpm">
+        ```bash
+        pnpm tauri dev
+        ```
+      </Tab>
+    </Tabs>
+  </Step>
+
+  <Step title="Continued Reading">
+    <Note>See [Desktop Apps](/apps/desktop-apps) for more information.</Note>
+  </Step>
+</Steps>
+
+## Testing the Tauri Mobile App
+
+We can use TestDriver and natural language to test our Tauri iOS app:
+
+<Steps>
+  <Step title="Run the Mobile App">
+    <Tabs>
+      <Tab title="npm">
+        ```bash
+        npm run tauri ios dev
+        ```
+      </Tab>
+      <Tab title="yarn">
+        ```bash
+        yarn tauri ios dev
+        ```
+      </Tab>
+      <Tab title="pnpm">
+        ```bash
+        pnpm tauri ios dev
+        ```
+      </Tab>
+    </Tabs>
+  </Step>
+
+  <Step title="Continued Reading">
+    <Note>See [Mobile Apps](/apps/mobile-apps) for more information.</Note>
+  </Step>
+</Steps>

package/docs/docs.json

@@ -64,7 +64,8 @@
               "apps/static-websites",
               "apps/desktop-apps",
               "apps/chrome-extensions",
-              "apps/mobile-apps"
+              "apps/mobile-apps",
+              "apps/tauri-apps"
             ]
           },
           {

package/docs/styles.css

@@ -53,4 +53,13 @@
   transform: translateX(-50%);
   justify-content: center;
   display: flex;
-}
+}
+
+img[src$=".svg"] {
+  background: transparent !important;
+}
+
+img[src$="https://tauri.app/favicon.svg"]
+{
+  opacity: 0.3;
+}

package/interfaces/cli/commands/generate.js

@@ -0,0 +1,3 @@
+const { createOclifCommand } = require("../utils/factory.js");
+
+module.exports = createOclifCommand("generate");

package/interfaces/cli/lib/base.js

@@ -55,7 +55,8 @@ class BaseCommand extends Command {
         `testdriverai-cli-${process.pid}.log`,
       );
 
-      console.log(`Log file created at: ${this.logFilePath}`);
+      console.log(`Log file: ${this.logFilePath}`);
+      console.log("");
       fs.writeFileSync(this.logFilePath, ""); // Initialize the log file
     }
 
@@ -68,11 +69,15 @@ class BaseCommand extends Command {
       );
     };
 
+    let isConnected = false;
+
     // Use pattern matching for log events, but skip log:Debug
     this.agent.emitter.on("log:*", (message) => {
       const event = this.agent.emitter.event;
 
       if (event === events.log.debug) return;
+
+      if (event === events.log.narration && isConnected) return;
       console.log(message);
     });
 
@@ -90,6 +95,7 @@ class BaseCommand extends Command {
 
     // Handle sandbox connection with pattern matching for subsequent events
     this.agent.emitter.on("sandbox:connected", () => {
+      isConnected = true;
       // Once sandbox is connected, send all log and error events to sandbox
       this.agent.emitter.on("log:*", (message) => {
         this.sendToSandbox(message);
@@ -134,6 +140,7 @@ class BaseCommand extends Command {
 
     // Handle show window events
     this.agent.emitter.on("show-window", async (url) => {
+      console.log("");
       console.log(`Live test execution: `);
       if (this.agent.config.CI) {
         let u = new URL(url);
@@ -178,7 +185,7 @@ class BaseCommand extends Command {
     // Prepare CLI args for the agent with all derived options
     const cliArgs = {
       command: this.id,
-      args: [filePath], // Pass the resolved file path as the first argument
+      args: filePath ? [filePath] : [], // Only pass file path if it exists
       options: {
         ...flags,
         resultFile:

package/interfaces/logger.js

@@ -44,7 +44,7 @@ class CustomTransport extends Transport {
 // responsible for rendering ai markdown output
 const { marked } = require("marked");
 const { markedTerminal } = require("marked-terminal");
-const { censorSensitiveData } = require("../agent/lib/censorship");
+const { censorSensitiveDataDeep } = require("../agent/lib/censorship");
 
 const { printf } = winston.format;
 
@@ -56,7 +56,7 @@ const logger = winston.createLogger({
   format: winston.format.combine(
     winston.format.splat(),
     winston.format((info) => {
-      info.message = censorSensitiveData(info.message);
+      info.message = censorSensitiveDataDeep(info.message);
       return info;
     })(),
     logFormat,
@@ -340,7 +340,7 @@ const createMarkdownLogger = (emitter) => {
 
     let diff = consoleOutput.replace(previousConsoleOutput, "");
     if (diff) {
-      diff = censorSensitiveData(diff);
+      diff = censorSensitiveDataDeep(diff);
       process.stdout.write(diff);
     }
   });
@@ -357,7 +357,7 @@ const createMarkdownLogger = (emitter) => {
     let diff = consoleOutput.replace(previousConsoleOutput, "");
 
     if (diff) {
-      diff = censorSensitiveData(diff);
+      diff = censorSensitiveDataDeep(diff);
       process.stdout.write(diff);
     }
     process.stdout.write("\n\n");

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "testdriverai",
-  "version": "6.0.32",
+  "version": "6.1.1-canary.613bfa3.0",
   "description": "Next generation autonomous AI agent for end-to-end testing of web & desktop",
   "main": "index.js",
   "bin": {

package/testdriver/examples/web/lifecycle/postrun.yaml

@@ -0,0 +1,7 @@
+version: 6.0.0
+steps:
+  - prompt: stop dashcam
+    commands:
+      - command: exec
+        lang: pwsh
+        code: dashcam -t 'Web Test Recording' -p

package/testdriver/examples/web/lifecycle/{provision.yaml → prerun.yaml}

@@ -2,6 +2,12 @@ version: 6.0.0
 steps:
   - prompt: launch chrome
     commands:
+      - command: exec
+        lang: pwsh
+        code: dashcam track --name=TestDriver --type=application --pattern="C:\Users\testdriver\Documents\testdriver.log"
+      - command: exec
+        lang: pwsh
+        code: dashcam start
       - command: exec
         lang: pwsh
         code: |

package/testdriver/lifecycle/postrun.yaml

@@ -1,8 +1,7 @@
-version: 5.1.1
-session: 67f00511acbd9ccac373edf7
+version: 6.0.0
 steps:
   - prompt: stop dashcam
     commands:
       - command: exec
         lang: pwsh
-        code: dashcam -t '${TD_THIS_FILE}' -p
+        code: dashcam -t 'Web Test Recording' -p

package/testdriver/lifecycle/prerun.yaml

@@ -1,7 +1,6 @@
 version: 6.0.0
-session: 67f00511acbd9ccac373edf7
 steps:
-  - prompt: start dashcam
+  - prompt: launch chrome
     commands:
       - command: exec
         lang: pwsh
@@ -9,3 +8,10 @@ steps:
       - command: exec
         lang: pwsh
         code: dashcam start
+      - command: exec
+        lang: pwsh
+        code: |
+          Start-Process "C:/Program Files/Google/Chrome/Application/chrome.exe" -ArgumentList "--start-maximized", "--guest", "https://testdriver-sandbox.vercel.app/login"
+      - command: wait-for-text
+        text: "TestDriver.ai Sandbox"
+        timeout: 60000

package/testdriver/lifecycle/provision.yaml

@@ -1,12 +0,0 @@
-version: 5.1.1
-session: 67f00511acbd9ccac373edf7
-steps:
-  - prompt: launch chrome
-    commands:
-      - command: exec
-        lang: pwsh
-        code: |
-          Start-Process "C:\Program Files\Google\Chrome\Application\chrome.exe" -ArgumentList "--start-maximized --disable-infobars --disable-fre --no-default-browser-check --no-first-run --guest --load-extension=$(pwd)/node_modules/dashcam-chrome/build", "https://testdriver-sandbox.vercel.app/"
-      - command: wait-for-text
-        text: ${TD_WEBSITE}
-        timeout: 60000