grepai-search-basics

This skill covers the fundamentals of semantic code search with GrepAI. Unlike traditional text search (grep, ripgrep), GrepAI searches by meaning : Output: By default, GrepAI returns 10 results. Adjust with --limit : Before searching, verify your index: Output: Describe the intent or behavior: Too short or generic: GrepAI understands natural language: Both work, but phrases often get better results: ❌ Problem: No results ✅ Solutions: ❌ Problem: Irrelevant results ✅ Solutions: ❌ Problem: Missing expected code ✅ Solutions: Successful basic search: - Learning GrepAI search - Performing basic code searches - Understanding semantic vs. text search - Interpreting search results - GrepAI initialized ( grepai init ) - Index created ( grepai watch ) - Embedding provider running (Ollama, etc.) - Finds exact text "authenticate" - Misses synonyms (login, signin, auth) - Returns all matches, unranked - Finds semantically similar code - Includes related concepts - Results ranked by relevance - Use English: Models are trained on English - Be specific: "JWT token validation" vs "validation" - Describe intent: What the code DOES, not what it's called - Use 3-7 words: Enough context, not too verbose - Iterate: Refine query based on results - Check index exists: grepai status - Run grepai watch if index is empty - Simplify query - Be more specific - Use different words - Check if code exists in...