From 9e5ffd2ccfa0223c7d0acbff52269b81cb7e6d55 Mon Sep 17 00:00:00 2001
From: Pavel Feldman <pavel.feldman@gmail.com>
Date: Tue, 27 May 2025 01:25:09 -0700
Subject: [PATCH] fix(cursor): allow enforcing images for cursor
 --image-responses=allow (#478)

Fixes https://github.com/microsoft/playwright-mcp/issues/449
---
 config.d.ts              | 4 ++--
 src/config.ts            | 8 ++------
 src/connection.ts        | 3 +--
 src/context.ts           | 9 +++++++++
 src/program.ts           | 2 +-
 src/tools/screenshot.ts  | 2 +-
 tests/screenshot.spec.ts | 4 ++--
 7 files changed, 18 insertions(+), 14 deletions(-)

diff --git a/config.d.ts b/config.d.ts
index 8b38e6d..1c010aa 100644
--- a/config.d.ts
+++ b/config.d.ts
@@ -117,7 +117,7 @@ export type Config = {
   };
 
   /**
-   * Do not send image responses to the client.
+   * Whether to send image responses to the client. Can be "allow", "omit", or "auto". Defaults to "auto", which sends images if the client can display them.
    */
-  noImageResponses?: boolean;
+  imageResponses?: 'allow' | 'omit' | 'auto';
 };
diff --git a/src/config.ts b/src/config.ts
index ef768da..047fdc0 100644
--- a/src/config.ts
+++ b/src/config.ts
@@ -38,7 +38,7 @@ export type CLIOptions = {
   host?: string;
   ignoreHttpsErrors?: boolean;
   isolated?: boolean;
-  imageResponses: boolean;
+  imageResponses?: 'allow' | 'omit' | 'auto';
   sandbox: boolean;
   outputDir?: string;
   port?: number;
@@ -188,13 +188,9 @@ export async function configFromCLIOptions(cliOptions: CLIOptions): Promise<Conf
     },
     saveTrace: cliOptions.saveTrace,
     outputDir: cliOptions.outputDir,
+    imageResponses: cliOptions.imageResponses,
   };
 
-  if (!cliOptions.imageResponses) {
-    // --no-image-responses was passed, disable image responses
-    result.noImageResponses = true;
-  }
-
   return result;
 }
 
diff --git a/src/connection.ts b/src/connection.ts
index 7083382..a315246 100644
--- a/src/connection.ts
+++ b/src/connection.ts
@@ -92,8 +92,7 @@ export class Connection {
     await new Promise<void>(resolve => {
       this.server.oninitialized = () => resolve();
     });
-    if (this.server.getClientVersion()?.name.includes('cursor'))
-      this.context.config.noImageResponses = true;
+    this.context.clientVersion = this.server.getClientVersion();
   }
 
   async close() {
diff --git a/src/context.ts b/src/context.ts
index d131609..8da2dbb 100644
--- a/src/context.ts
+++ b/src/context.ts
@@ -48,12 +48,21 @@ export class Context {
   private _modalStates: (ModalState & { tab: Tab })[] = [];
   private _pendingAction: PendingAction | undefined;
   private _downloads: { download: playwright.Download, finished: boolean, outputFile: string }[] = [];
+  clientVersion: { name: string; version: string; } | undefined;
 
   constructor(tools: Tool[], config: FullConfig) {
     this.tools = tools;
     this.config = config;
   }
 
+  clientSupportsImages(): boolean {
+    if (this.config.imageResponses === 'allow')
+      return true;
+    if (this.config.imageResponses === 'omit')
+      return false;
+    return !this.clientVersion?.name.includes('cursor');
+  }
+
   modalStates(): ModalState[] {
     return this._modalStates;
   }
diff --git a/src/program.ts b/src/program.ts
index 1f49736..8638342 100644
--- a/src/program.ts
+++ b/src/program.ts
@@ -40,7 +40,7 @@ program
     .option('--host <host>', 'host to bind server to. Default is localhost. Use 0.0.0.0 to bind to all interfaces.')
     .option('--ignore-https-errors', 'ignore https errors')
     .option('--isolated', 'keep the browser profile in memory, do not save it to disk.')
-    .option('--no-image-responses', 'do not send image responses to the client.')
+    .option('--image-responses <mode>', 'whether to send image responses to the client. Can be "allow", "omit", or "auto". Defaults to "auto", which sends images if the client can display them.')
     .option('--no-sandbox', 'disable the sandbox for all process types that are normally sandboxed.')
     .option('--output-dir <path>', 'path to the directory for output files.')
     .option('--port <port>', 'port to listen on for SSE transport.')
diff --git a/src/tools/screenshot.ts b/src/tools/screenshot.ts
index 3317103..439d79a 100644
--- a/src/tools/screenshot.ts
+++ b/src/tools/screenshot.ts
@@ -64,7 +64,7 @@ const screenshot = defineTool({
     else
       code.push(`await page.screenshot(${javascript.formatObject(options)});`);
 
-    const includeBase64 = !context.config.noImageResponses;
+    const includeBase64 = context.clientSupportsImages();
     const action = async () => {
       const screenshot = locator ? await locator.screenshot(options) : await tab.page.screenshot(options);
       return {
diff --git a/tests/screenshot.spec.ts b/tests/screenshot.spec.ts
index b1d73f6..d771359 100644
--- a/tests/screenshot.spec.ts
+++ b/tests/screenshot.spec.ts
@@ -172,12 +172,12 @@ test('browser_take_screenshot (filename: "output.jpeg")', async ({ startClient,
   expect(files[0]).toMatch(/^output\.jpeg$/);
 });
 
-test('browser_take_screenshot (noImageResponses)', async ({ startClient, server }, testInfo) => {
+test('browser_take_screenshot (imageResponses=omit)', async ({ startClient, server }, testInfo) => {
   const outputDir = testInfo.outputPath('output');
   const client = await startClient({
     config: {
       outputDir,
-      noImageResponses: true,
+      imageResponses: 'omit',
     },
   });