Overview

dissolution is a local semantic search layer for code and docs.

Core flow

1. Start the daemon once:

dissolve start

2. Add one or more directories:

dissolve add .
dissolve add ~/projects/another-repo

3. Search from CLI or UI:

dissolve search "how auth token refresh works"
dissolve web

What gets indexed?

• Code, markdown, text, config files
• Any file not ignored by .gitignore / .ignore
• Per-directory SQLite DB at <path>/.dissolution/index.db by default

Why SQLite?

• Single binary, no background DB dependency
• Good enough KNN via sqlite-vec
• Easy backup/migration with plain files