Skip to content

Latest commit

 

History

History
161 lines (118 loc) · 6.03 KB

SHAREX_SETUP.md

File metadata and controls

161 lines (118 loc) · 6.03 KB

ShareX Integration Guide

This guide explains how to set up ShareX to work with AIPromptBridge for image processing tasks like OCR, translation, and description.

Note: AIPromptBridge now includes a built-in Screen Snipping feature (Ctrl+Alt+X) which provides similar functionality without requiring external tools. This ShareX integration is intended for advanced users who need ShareX's specific workflow capabilities or custom endpoints.

Demo

ShareX_AI-Bridge-Demo.mp4

Prerequisites

  • Latest version of ShareX installed
  • AIPromptBridge running with Flask endpoints enabled:
    1. Open config.ini (Right-click tray icon -> Edit config.ini)
    2. Set flask_endpoints_enabled = true under [config] (it is disabled by default)
    3. Restart AIPromptBridge

Step 1: Create Custom Uploader

  1. Open ShareX
  2. Go to Custom uploader settings
  3. Click New to create a new uploader

Step 2: Configure the Uploader

Fill in the following settings:

Setting Value
Name AIPromptBridge - OCR (or any name)
Destination type Image uploader
Method POST
Request URL http://127.0.0.1:5000/ocr
Body Form data (multipart/form-data)
File form name image
URL {response}

Custom uploader settings

Available Endpoints

AIPromptBridge comes with these default endpoints (configurable in prompts.json):

Endpoint Purpose
/ocr Extract text from image
/ocr_translate Extract and translate text (use ?lang= parameter)
/translate Translate image text to English
/describe Describe image content
/summarize Summarize image content
/code Extract and format code from image
/explain Analyze and explain image content
/latex Convert equations to LaTeX
/markdown Convert content to Markdown
/proofread Extract and proofread text

Custom Endpoints

You can add your own endpoints by editing prompts.json (Right-click tray -> Settings -> Endpoints Tab). The endpoints are defined in the endpoints section of the JSON configuration.

Query Parameters

Add these to the URL for additional behavior:

Parameter Example Effect
?show=yes /ocr?show=yes Open result in chat window for follow-up
?show=no /ocr?show=no Return text only (default)
?lang=XX /ocr_translate?lang=Japanese Target language for translation
?prompt=... /ocr?prompt=Extract+only+numbers Override the endpoint prompt
?model=... /ocr?model=gemini-2.5-pro Override the model
?provider=... /ocr?provider=openrouter Override the provider

Examples:

  • http://127.0.0.1:5000/ocr?show=yes - OCR with chat follow-up
  • http://127.0.0.1:5000/ocr_translate?lang=English - Translate to English
  • http://127.0.0.1:5000/ocr_translate?lang=Indonesian - Translate to Indonesian
  • http://127.0.0.1:5000/describe?show=yes&model=gemini-2.5-pro - Describe with specific model

Step 3: Create Hotkey Workflow

  1. Go to Hotkey settings (or right-click tray → Hotkey settings)
  2. Click Add... to create a new hotkey
  3. Set:
    • Task: Capture region (or Capture active window, etc.)
    • Hotkey: Your preferred key combination

Hotkey setup

Step 4: Configure Task Settings

  1. Click the Gear icon ⚙️ next to your new hotkey
  2. Configure Override after capture tasks:
    • Check Upload image to host
  3. Configure Override after upload tasks:
    • Check Copy URL to clipboard
  4. Configure Override destinations:
    • Check Image uploader
    • Select Custom image uploader
  5. Configure Override default custom uploader:
    • Check and select your AIPromptBridge - OCR uploader

Task settings

Usage

  1. Press your configured hotkey
  2. Select a region on screen (or capture window/fullscreen)
  3. Wait for AIPromptBridge to process the image
  4. The extracted text is now in your clipboard - paste anywhere!

Quick Access via Tray

You can also access workflows by:

  1. Right-click ShareX tray icon
  2. Go to Workflows
  3. Select your workflow

Troubleshooting

"Connection refused" Error

  • Make sure AIPromptBridge is running (check system tray for the icon)
  • Check if Flask endpoints are enabled: Ensure flask_endpoints_enabled = true in config.ini
  • Check the port matches your config (default: 5000)
  • Try opening http://127.0.0.1:5000/ in your browser

Empty Response

  • Check AIPromptBridge console for errors (right-click tray → Show Console)
  • Verify your API keys are configured in config.ini
  • Make sure to use models that supports vision

Slow Processing

  • Use a faster model (e.g., gemini-2.5-flash instead of gemini-2.5-pro)
  • Disable thinking mode: Set thinking_enabled = false in config.ini
  • Use non-reasoning models for simple tasks

Example Workflows

Create multiple custom uploaders for different tasks:

Workflow URL Use Case
Quick OCR /ocr Fast text extraction
OCR + Chat /ocr?show=yes Text with follow-up questions
Translate to English /ocr_translate?lang=English Any language → English
Translate to Japanese /ocr_translate?lang=Japanese Any language → Japanese
Describe Image /describe?show=yes Image description with chat
Extract Code /code Screenshot to code
Explain Code /explain_code?show=yes Code explanation

Tips

  • Use ?show=yes when you want to ask follow-up questions about the image
  • Create separate hotkeys for different workflows (e.g., Ctrl+1 for OCR, Ctrl+2 for translate)
  • The response is plain text - it works with any application that accepts paste
  • Add your own custom endpoints in prompts.json for specialized prompts
  • Use the {lang} placeholder in custom endpoints for dynamic language selection