Skip to content

Context Awareness

DevLingo automatically detects the app you’re currently using and adjusts the AI learning card’s style, examples, and tone accordingly. This means the same word may get different explanations in different apps.

Style: Technical, code examples, API-style usage

When looking up vocabulary in a code editor or terminal, DevLingo provides:

  • Code snippet examples: Shows how to use the term in actual code
  • Related APIs or libraries: Links to official documentation
  • Technical terminology: Precise computer science expressions
  • Common pitfalls: Language traps developers frequently fall into

Example: Looking up “deprecate”

  • In Xcode: Explained as API deprecation, including @deprecated annotation examples
  • In Slack: Explained as “phasing out,” with more casual usage

Style: Casual, concise, emoji-friendly

When looking up in chat apps, results will:

  • Be brief: Avoids lengthy explanations
  • Use friendly expressions: Emoji and abbreviation support
  • Focus on team collaboration context: Emphasizes usage in team communication
  • Be real-time: Explains how to express things quickly in discussions

Example: Looking up “refactor”

  • In Slack: Focuses on code improvement with a casual tone
  • In Xcode: Detailed refactoring technical patterns and tools

Style: Academic, detailed, grammatically deep

When browsing web pages and documentation:

  • Grammar analysis: Detailed grammatical structure explanations
  • Etymology and evolution: Historical background of vocabulary
  • Synonym differentiation: Fine-grained meaning distinctions
  • Formal register: Academic and professional usage

Example: Looking up “implement”

  • In Safari: Deep dive into multiple meanings of “implement,” including legal and business usage
  • In VS Code: Emphasizes the technical aspects of code implementation

Style: Business formal, email-appropriate

When looking up in email apps:

  • Formal tone: Appropriate expressions for business communication
  • Email conventions: How to use terms properly in emails
  • Polite expressions: Cross-cultural business communication sensitivity
  • Signature-friendly: Expressions suitable for email closings

Example: Looking up “proceed”

  • In Mail: Emphasizes email body and action step descriptions
  • In Slack: Casual “continue” or “move forward”

Style: Code review, PR/Issue conventions

When looking up in GitHub web pages (Issues, PRs, Discussions):

  • Code review terminology: Standard expressions for PR comments
  • Issue descriptions: How to clearly describe problems and requirements
  • Conventions and agreements: Common language in the open-source community
  • Documentation suggestions: Best practices for Markdown and code documentation

DevLingo uses the macOS API (NSWorkspace.shared.frontmostApplication.bundleIdentifier) to detect the current app, then:

  1. Identifies app type: Xcode, Slack, Safari, etc.
  2. Categorizes app context: code, chat, documentation, email, etc.
  3. Customizes AI prompt: Passes context labels to the Claude API
  4. Adjusts response format: Code editors get code blocks; chat gets concise formats

The same word “leverage” across different apps:

Xcode (code editor)

leverage (verb, technical)
Usage in code: Developers leverage APIs to...
Example: We leverage the Observer pattern for state management.
Synonyms: utilize, employ, use

Slack (chat)

leverage - to utilize, make the most of
Casual: "We should leverage our team's expertise on this"
Quick tip: Use in discussions about teamwork or resources

Mail (email)

leverage (formal, business)
Professional: "We aim to leverage existing partnerships"
Register: Formal, suitable for business communication

Future versions will support custom app classification. If DevLingo misidentifies your app, you’ll be able to:

  1. Add custom app rules in settings
  2. Assign a context (code, chat, documentation, etc.) to a specific app
  3. Save preferences for automatic application in future lookups
  • App detection data is not logged or uploaded
  • Only the lookup vocabulary is sent to the Claude API
  • App context labels improve AI responses and are not used for user tracking

Think the context style for a particular app isn’t accurate enough? We welcome feedback:

We continuously optimize context prompts for each app based on user feedback.