grepai-config-reference

This skill serves as the primary technical reference for managing GrepAI configurations through the .grepai/config.yaml file. It is designed for software engineers, DevOps professionals, and AI agent developers who need to fine-tune semantic code search environments for various project scales, from local personal projects to complex team-based architectures. The skill covers the entire schema, enabling users to switch between local embedding providers like Ollama and cloud-based alternatives like OpenAI, or adjust storage backends such as GOB, PostgreSQL, and Qdrant. By leveraging this reference, users can optimize indexing performance, manage file watching with precision, and apply search boosting strategies to prioritize relevant codebase areas like core logic over noise like test directories or dependency folders.

• Complete schema breakdown of the .grepai/config.yaml file.

• Configuration instructions for embedder providers including Ollama, OpenAI, and LM Studio.

• Setup details for storage backends like PostgreSQL (with pgvector) and Qdrant.

• Tuning parameters for code chunking, token sizing, and overlap to balance context vs. performance.

• Scoring and ranking configuration for search result boosting (penalties and bonuses).

• Advanced trace configuration for call graph analysis using tree-sitter AST parsing.

• Ignore pattern management to exclude irrelevant files like node_modules, build artifacts, or generated code.

• Environment variable support for sensitive API key injection.

• Recommended for configuring local-first, privacy-conscious AI agent setups.

• Provides specific configuration templates for varying use cases: small projects, large-scale codebases, team environments, and high-privacy scenarios.

• Essential for troubleshooting index synchronization, search relevancy, and embedding vector dimension mismatches.

• The schema supports dynamic path exclusions and custom regex patterns for targeted search scopes.

• Always verify your .grepai/config.yaml syntax to ensure the agent correctly parses your vector storage backend and indexing requirements.