index

omnisess index

Bulk-populate the transcript SQLite cache used by omnisess stats.

Usage

omnisess index --all [--full] [--rebuild]

• --all is required (other modes are not yet implemented).
• --full captures tool-use arguments and result payloads. This is privacy-sensitive: Bash command strings and Write/Edit file content may contain secrets. Treat the cache file as user-private data.
• --rebuild drops existing rows for each session before re-indexing. Use this after a schema change or to recover from a corrupt cache.

The index file lives at os.UserCacheDir() + /omnisess/index.sqlite by default; override via OMNISESS_INDEX_PATH.

When you need this

You don't, strictly — omnisess stats lazy-indexes on first query. But:

• One bulk run up front is much faster than indexing during interactive use.
• --full only makes per-session data available if it was captured at index time. omnisess stats --full cannot retroactively populate arguments/results that weren't recorded.

Instructions

1. Check that the omnisess binary is available:

if ! command -v omnisess &>/dev/null; then
  echo "omnisess not found. Install it with:"
  echo "  go install github.com/psacc/omnisess@latest"
  exit 1
fi

2. Run the command, passing through all arguments:

omnisess index "$@"

Examples

# Seed the cache with metadata-only rows (default)
omnisess index --all

# Capture full arguments and result payloads (privacy-sensitive)
omnisess index --all --full

# Drop existing rows and re-index
omnisess index --all --rebuild