Update ScreenOCR extension with cross-extension conventions

[chrismessina]

Description

This PR implements cross-extension support for ScreenOCR, allowing other Raycast extensions to programmatically call OCR functionality and receive structured data via callbacks. It also includes several code quality improvements and dependency updates.

🌟 Key Changes

1. Cross-Extension Integration

   • Implemented raycast-cross-extension-conventions.
   • Added raycast-cross-extension as a dependency.
   • Updated package.json with necessary metadata to act as a cross-extension provider.
   • Added a new guide: docs/cross-extension-usage.md to help other developers integrate with ScreenOCR.

2. Enhanced Type Safety & Readability

   • Introduced dedicated OCRResult and LaunchContext types for robust callback handling.
   • Refactored recognize-text.tsx and detect-barcode.tsx to use early return patterns, significantly cleaning up nested logic.
   • Updated error handling to leverage @raycast/utils while maintaining support for user-defined sound and notification preferences.

3. Maintenance & Dependency Updates

   • Upgraded @raycast/api to ^1.104.1.
   • Updated dev dependencies including typescript (^5.9.3) and @types packages for React/Node.
   • Refined CHANGELOG.md and README.md to reflect these updates.

Checklist

• I read the extension guidelines
• I read the documentation about publishing
• I ran npm run build and tested this distribution build in Raycast
• I checked that files in the assets folder are used by the extension itself
• I checked that assets used by the README are placed outside of the metadata folder

[raycastbot]

Thank you for your contribution! 🎉

🔔 @neo773 @anwarulislam @xilopaint @pernielsentikaer you might want to have a look.

You can use this guide to learn how to check out the Pull Request locally in order to test it.

📋 Quick checkout commands

BRANCH="ext/screenocr"
FORK_URL="https://github.com/chrismessina/extensions.git"
EXTENSION_NAME="screenocr"
REPO_NAME="extensions"

git clone -n --depth=1 --filter=tree:0 -b $BRANCH $FORK_URL
cd $REPO_NAME
git sparse-checkout set --no-cone "extensions/$EXTENSION_NAME"
git checkout
cd "extensions/$EXTENSION_NAME"
npm install && npm run dev

Due to our current reduced availability, the initial review may take up to 10-15 business days.

[neo773]

Hi,

I just had a quick look at this on my phone and the code looks good. Could you remove the redundant, unnecessary code comment and also use early returns instead of nested if/else?

Thanks for the contribution.

[chrismessina]

@greptile could you take a look?

[greptile-apps]

I'll review the screenocr extension update. Let me look at the changes:

Summary

This PR adds cross-extension support to the ScreenOCR extension, allowing other Raycast extensions to call ScreenOCR programmatically and receive OCR results via callbacks.

Key Changes

✅ Good Implementation

1. Proper Cross-Extension Integration

   • Added raycast-cross-extension dependency
   • Implemented callback mechanism in recognize-text.tsx
   • Added comprehensive documentation

2. Backwards Compatibility

   • Extension still works as before when called directly
   • Only uses callback when launched from another extension

3. Error Handling

   • Proper error propagation to callback consumers
   • Handles both "no text detected" and general failures

4. Documentation

   • Good cross-extension usage documentation
   • Clear examples for consumer setup

🔍 Areas for Improvement

1. Type Safety in recognize-text.tsx

The callback data structure could be more strongly typed:

import { Clipboard, closeMainWindow, LaunchProps } from "@raycast/api";
import { LaunchOptions, callbackLaunchCommand } from "raycast-cross-extension";
import { recognizeText, showSuccessToast, showFailureToast } from "./utils";

type OCRResult = {
  text: string | null;
  error?: string;
};

type LaunchContext = {
  callbackLaunchOptions?: LaunchOptions;
};

export default async function command({
  launchContext,
}: LaunchProps<{ launchContext?: LaunchContext }>) {
  await closeMainWindow();

  const callbackOptions = launchContext?.callbackLaunchOptions;

  try {
    const recognizedText = await recognizeText();

    if (!recognizedText) {
      await showFailureToast("No text detected");
      // If there's a callback, send empty result back
      if (callbackOptions) {
        await callbackLaunchCommand(callbackOptions, {
          text: null,
          error: "No text detected",
        } as OCRResult);
      }
      return;
    }

    // If called from another extension with callback, send result back
    if (callbackOptions) {
      await callbackLaunchCommand(callbackOptions, { text: recognizedText } as OCRResult);
    } else {
      // Standard behavior: copy to clipboard
      await Clipboard.copy(recognizedText);
      await showSuccessToast("Copied text to clipboard");
    }
  } catch (e) {
    console.error(e);
    await showFailureToast("Failed detecting text");
    // If there's a callback, send error back
    if (callbackOptions) {
      await callbackLaunchCommand(callbackOptions, {
        text: null,
        error: e instanceof Error ? e.message : "Failed detecting text",
      } as OCRResult);
    }
  }
}

2. Cross-Extension Badge

   • ✅ Added the cross-extension badge to README
   • ✅ Updated metadata correctly

3. Documentation Structure

   • The cross-extension docs are well-structured
   • Example code is clear and practical

Overall Assessment

This is a solid implementation of cross-extension support that:

• Maintains backward compatibility
• Provides clear documentation
• Implements proper error handling
• Uses the cross-extension conventions correctly

The PR successfully transforms ScreenOCR into a provider extension while keeping all existing functionality intact. Great work! 🎉

[greptile-apps]

Greptile Summary

Implemented cross-extension support for ScreenOCR following the raycast-cross-extension-conventions, enabling other extensions to programmatically trigger OCR and receive structured results via callbacks.

• Added OCRResult and LaunchContext types with satisfies validation for type-safe callback handling
• Enhanced recognize-text.tsx to support callback-based cross-extension communication with proper error handling
• Refactored code with early return patterns across multiple files for improved readability
• Modernized error handling using @raycast/utils while maintaining user preference support for toast notifications and sounds
• Added comprehensive documentation in docs/cross-extension-usage.md
• Updated dependencies to latest versions (@raycast/api ^1.104.1, @raycast/utils ^2.2.2)
• CHANGELOG follows proper formatting with {PR_MERGE_DATE} placeholder at the top

Confidence Score: 5/5

• This PR is safe to merge with no issues found
• The implementation follows Raycast conventions properly, with correct cross-extension setup, excellent type safety using satisfies, proper error handling, and well-structured code refactoring. All changes are clean, focused, and maintain backward compatibility.
• No files require special attention

Important Files Changed

Filename	Overview
extensions/screenocr/src/recognize-text.tsx	Implemented cross-extension callback support with proper error handling and early returns for improved readability
extensions/screenocr/src/types.ts	Added OCRResult and LaunchContext types for type-safe cross-extension communication
extensions/screenocr/src/utils.ts	Refactored toast helpers with early returns and integrated @raycast/utils for better error handling
extensions/screenocr/package.json	Added raycast-cross-extension dependency and updated @raycast packages to latest versions
extensions/screenocr/docs/cross-extension-usage.md	New documentation file explaining how other extensions can consume ScreenOCR's functionality