Skip to main content

Mask Feature Overview

1. Region-Based Overlay Translation​

You can select only the region you want to translate. Mask processes the selected area instead of the full screen, reducing irrelevant text noise.

Typical scenarios:

  • Dialogue boxes in foreign-language games
  • Text bubbles in comics
  • Paragraph-level translation in presentations

2. Context-Aware Translation​

Mask does more than plain OCR. It uses scene context to improve readability and meaning in visually complex content.

Typical scenarios:

  • Content with charts and diagrams
  • Handwritten notes or non-standard layout
  • Stylized fonts and mixed visual elements

3. Privacy-First Experience​

Mask is designed with privacy first. Core recognition can run locally, making it suitable for privacy-sensitive workflows.

Typical scenarios:

  • Work-related document reading
  • Personal content with sensitive data
  • Cases where screenshot content should not be uploaded

4. Multilingual Reading Support​

Mask is built for cross-language reading and long-form comparison.

Typical scenarios:

  • International documentation reading
  • Bilingual learning
  • Cross-language information retrieval

5. Multi-Provider Model Access​

Mask supports multiple model providers, so you can switch based on cost, speed, and quality.

Common usage patterns:

  • Daily reading: prioritize lower-cost models
  • Difficult content: switch to higher-quality models
  • Sensitive content: use local or self-hosted inference

6. Local Configuration and History​

API keys, conversation history, and preferences are stored locally by default, which helps with long-term usage and context continuity.

Related docs: