Indexing and Daemon

Which Mode Should I Use?

Scenario	Recommended
Quick one-off searches	run search/read/definition directly (auto bootstrap)
Active coding session	cgrep daemon start while coding, then cgrep daemon stop
One-time async prebuild	cgrep index --background
Semantic/hybrid experiments	cgrep index --embeddings precompute (experimental)

Core Commands

# Build or refresh index
cgrep index

# Force full rebuild
cgrep index --force

# One-shot background build
cgrep index --background

# Daemon lifecycle
cgrep daemon start
cgrep daemon status
cgrep daemon stop

# Read readiness quickly
cgrep status

Large Repository Tips

# Lower pressure settings for high-churn repos
cgrep daemon start --debounce 30 --min-interval 180 --max-batch-delay 240

Defaults are already tuned for background operation; adjust only if needed.

Notes

• Index files are stored in .cgrep/.
• Ignore files (.gitignore, .ignore) are respected by default.
• --include-ignored disables ignore filtering.
• --include-path <path> lets you include selected ignored paths.
• Daemon is event-driven; without file changes it stays idle.