Skip to main content
Bob is a powerful translation and OCR software for macOS and iOS. Through Laozhang API, you can integrate various advanced AI models into Bob for translation, polishing, summarizing and other functions.

Quick Integration

Step 1: Install Bob

  1. Visit Bob Official Website to download and install
  2. Or install via Homebrew:
    brew install bob
    

Step 2: Install OpenAI Plugin

  1. Open Bob application
  2. Click the menu bar icon, select “Preferences”
  3. Go to “Plugins” tab
  4. Search for “OpenAI Translate” plugin
  5. Click “Install” button

Step 3: Configure Laozhang API

1

Open Plugin Settings

Find “OpenAI Translate” plugin in the plugin list, click the settings icon
2

Fill In API Configuration

Configure the following information:
  • API Base URL: https://api.laozhang.ai
  • API Key: Enter your Laozhang API key
  • Model: Select the model you need (e.g., gpt-4-turbo)
3

Test Configuration

Click “Test” button to ensure connection is successful
4

Save Settings

Click “Save” button to complete configuration
Obtaining API KeyVisit Laozhang API Console to create and obtain your API key.
Choose appropriate models based on different translation scenarios:
Translation ScenarioRecommended ModelReason
Daily Translationgpt-3.5-turboFast response, economical
Professional Documentsgpt-4-turboHigh accuracy, strong domain knowledge
Literary Worksclaude-sonnet-4Strong literary flair
Technical Documentsgpt-4-turboPrecise technical terminology
Simultaneous Interpretationgpt-3.5-turboUltra-fast response
Cost vs. Quality Trade-off
  • Daily use: Recommend gpt-3.5-turbo, cost-effective
  • Important documents: Recommend gpt-4-turbo or claude-sonnet-4, higher quality
  • Large batches: Consider using gpt-3.5-turbo combined with manual proofreading

Core Features

1. Text Translation

Select any text, use shortcuts to translate:
  • Screenshot Translation: ⌥ + A
  • Selection Translation: ⌥ + D
  • Input Translation: ⌥ + S

2. OCR Text Recognition

Extract text from images:
  • Support multiple languages
  • High recognition accuracy
  • Direct translation of recognition results

3. Text Polishing

Optimize your text:
  • Grammar check
  • Expression optimization
  • Style adjustment
  • Tone conversion

4. Text Summarization

Quickly summarize long text:
  • Extract key points
  • Generate abstracts
  • Refine core ideas

Practical Examples

Scenario 1: Web Translation

Browsing foreign language websites:
  1. Select the text you want to translate
  2. Use shortcut ⌥ + D
  3. View translation result in popup

Scenario 2: Document Translation

Translating professional documents:
  1. Copy document text
  2. Use shortcut ⌥ + S
  3. Input or paste text
  4. Get high-quality translation

Scenario 3: Paper Reading

Reading English papers:
  1. Screenshot paper paragraphs
  2. Use shortcut ⌥ + A
  3. OCR extracts text
  4. Automatic translation display

Scenario 4: Email Writing

Writing English emails:
  1. Write Chinese draft
  2. Use translation feature
  3. Get English version
  4. Use polishing feature for optimization

Advanced Configuration

Custom Prompts

Customize prompts for translation:
You are a professional translator proficient in {source language} and {target language}.

Task: Translate the following text to {target language}
Requirements:
1. Accurate and fluent translation
2. Maintain original style and tone
3. Handle proper nouns appropriately
4. Preserve formatting

Text to translate:
{text}

Multi-model Configuration

Configure different models for different scenarios:
ScenarioModelTemperatureMax Tokens
General Translationgpt-3.5-turbo0.32000
Literary Translationclaude-sonnet-40.73000
Technical Translationgpt-4-turbo0.22000
Text Polishinggpt-4-turbo0.52000

Shortcut Customization

Customize shortcuts as needed:
FunctionDefault ShortcutRecommended Custom
Screenshot Translation⌥ + ASet to your preferred
Selection Translation⌥ + DSet to your preferred
Input Translation⌥ + SSet to your preferred
Polishing FeatureNone⌥ + P (recommended)

Troubleshooting

Connection Failed

Problem: Unable to connect to API Solutions:
  1. Check if API Base URL is correct: https://api.laozhang.ai
  2. Verify API Key validity
  3. Check network connection
  4. Confirm firewall settings

Translation Quality Poor

Problem: Translation results don’t meet expectations Solutions:
  1. Try using a more advanced model (e.g., switch from GPT-3.5 to GPT-4)
  2. Adjust temperature parameter (lower = more accurate, higher = more diverse)
  3. Optimize prompt template
  4. Provide more context information

Slow Response

Problem: Translation speed is slow Solutions:
  1. Switch to a faster model (e.g., gpt-3.5-turbo)
  2. Reduce max_tokens setting
  3. Check network speed
  4. Try during off-peak hours

High Cost

Problem: Usage costs are too high Solutions:
  1. Use gpt-3.5-turbo for daily translation
  2. Use advanced models only for important content
  3. Set reasonable max_tokens limits
  4. Regularly check usage

Best Practices

1. Choose Models Wisely

Choose appropriate models based on scenarios:
Daily communication → gpt-3.5-turbo (fast and economical)

Professional documents → gpt-4-turbo (highly accurate)

Literary works → claude-sonnet-4 (strong literary flair)

2. Optimize Prompts

Provide clear translation requirements:
❌ Bad example: "Translate this text"

✅ Good example: "Please translate the following technical documentation text from English to Chinese, pay attention to:
1. Keep technical terminology professional
2. Maintain original paragraph structure
3. Preserve code blocks and formatted content
4. Ensure text is fluent and natural"

3. Context Management

Provide necessary context information:
  • Document type (paper, email, chat, etc.)
  • Domain background (technical, medical, legal, etc.)
  • Target audience (professional, general public, etc.)
  • Style requirements (formal, casual, etc.)

4. Result Verification

Check translation quality:
  • Verify terminology accuracy
  • Check logic coherence
  • Confirm style appropriateness
  • Proofread for typos

Usage Tips

Efficient Workflow

1. Selection/Screenshot

2. Automatic Recognition/Translation

3. Check Results

4. (If needed) Secondary Polishing

5. Copy and Use

Batch Processing

Handle large amounts of text:
  1. Split large text into appropriate blocks
  2. Process each block sequentially
  3. Merge results
  4. Perform overall proofreading

Quality Control

Ensure translation quality:
  1. Initial Translation: Use GPT-3.5 for rough translation
  2. Accuracy Check: Important parts use GPT-4 for retranslation
  3. Style Polishing: Use Claude for style optimization
  4. Final Review: Manual review and adjustment

Plugin Extensions

Plugin NameFunctionUse Case
DeepL TranslateProfessional translationOfficial translation alternative
Google TranslateQuick translationDaily translation supplement
Youdao TranslateChinese-English translationDomestic user-friendly
OpenAI PolishingText polishingContent optimization

Plugin Combination Strategies

Bob OpenAI Plugin (Primary)

DeepL Plugin (Comparison)

Manual Proofreading (Quality Assurance)

Cost Optimization

Usage Strategy

Document TypeStrategyExpected Cost
Chat MessagesGPT-3.5, no optimizationVery Low
General DocumentsGPT-3.5, batch processingLow
Important DocumentsGPT-4, secondary proofreadingMedium
Contract/LegalGPT-4 + Manual ReviewHigh

Budget Control

Set budget control:
  1. Daily Cap: Limit daily maximum usage
  2. Priority Allocation: Allocate quota by importance
  3. Monitor Usage: Regularly check usage
  4. Optimize Model Selection: Adjust model strategy based on actual needs

Further Learning

Official Resources

Learning Resources

Community Support

Need more help? Please visit Laozhang API Official Website or contact our support team.
I