scan

Reads a folder recursively and records every supported media file in the database. Run this first: everything else reads from what it creates.

videre scan ~/Photos                              # scan into the default database
videre scan                                       # same, using the folder from `videre config set path`
videre scan ~/Photos --similar                    # also fingerprint images/videos for near-duplicate detection
videre scan ~/Photos --db ~/photos.db             # write to a specific database instead
videre scan ~/Photos --output                     # write JSONL to ~/.videre/hashes.jsonl instead of SQLite
videre scan ~/Photos --output out.jsonl           # write JSONL to a specific file
videre scan ~/Photos --retry-incomplete           # only files an earlier scan didn't finish
videre scan ~/Photos --silent                     # no progress output
videre scan ~/Photos --json                       # print one JSON summary object instead
videre scan ~/Photos --type video                 # only videos
videre scan ~/Photos --ext heic,mov               # only these formats
videre scan ~/Photos --path ~/Photos/2024         # only this subfolder

:::tip These filters work the same way across commands, and combine. See scoping a run. :::

Re-running is safe and idempotent, since existing entries are updated in place.

:::note --output and --db cannot be combined. A bare --output must come after the folder, or it swallows the folder as its value.

--output-sqlite still works as an alias for --db. It was the original name, from when JSONL and SQLite were peer output formats rather than one destination and one opt-out. Existing scripts do not need changing. :::

What it records

For every file: its content hash, size, timestamps, extension, detected type, and EXIF where present (date taken, GPS, dimensions). Nothing is opened for decoding unless you pass --similar.

It does not prepare search or detect faces. Those are videre embed and videre faces, run separately and much slower.