Search across your project using multiple powerful methods—from fast text matching to AI-powered semantic search.
Search Methods
Embedr offers three search approaches:
| Method | Description | Use Case |
|---|---|---|
| Grep Search | Fast text matching using ripgrep | Finding exact strings, function names, variable references |
| Semantic Search | AI-powered conceptual search | Finding code by description, understanding, or intent |
| Hybrid Search | Combined text + semantic search | Best overall accuracy for complex queries |
Project Content Search (Grep)
Use the integrated grep tool for fast text searches:
- Via AI Agent: "Search for
WiFi.beginin my project" - Results include file path, line number, and matches
Options
- Case sensitivity
- Regex vs fixed strings
- Include hidden files
- Respect or ignore .gitignore
- Limit search paths or globs
Semantic Code Search
Semantic search requires Codebase Indexing to be enabled.
Open Search Palette
Use keyboard shortcuts:
- macOS: ⌘+Shift+F or ⌘+P
- Windows: Ctrl+Shift+F or Ctrl+P
Search Modes
- Hybrid (recommended): Combines semantic and text search
- Semantic: Pure AI-powered search—understands concepts
- Quick: Fast text-only search using ripgrep
Example Queries
- "function that initializes the display"
- "code handling button debounce"
- "where sensors are calibrated"
Engine
- Uses ripgrep (
rg) when available; falls back to a Node.js scanner for portability - Semantic search uses AI embeddings for concept matching
Tips
- Enable codebase indexing for semantic search capabilities
- Narrow your globs for large repos
- Combine grep for exact matches with semantic for understanding