Search & Discovery

File search, metadata search, content search, and ID3 tag queries

Overview

copyparty provides powerful search capabilities: find files by name/path, size, date, content hash, or metadata tags like artist and title.

File Search by Content

Find duplicate files by dropping them into the browser.

How It Works

1. Drag files into the Search dropzone (or toggle [🔎] on mobile)
2. Files are hashed client-side
3. Hash sent to server
4. Server checks database for matching content
5. Results show where identical files exist

Requires -e2dsa to index file hashes. Files that match go into [ok], unknown files into [ng].

Wark Search

Search by file identifier (wark/checksum) directly:

raw: w = kFpDiztbZc8Z1Lzi

Enter this in the [🔎] Search tab → raw field.

Metadata Search

Search by file attributes and media tags.

Enable Indexing

# Enable file indexing and tag scanning
copyparty -e2dsa -e2ts

[global]
  e2dsa  # index all files (size, date, hash)
  e2ts   # index audio/video tags

Search Interface

Click [🔎] in the top navigation to open the search tab.

Search Fields

Basic

Path

• Space-separated AND search
• - prefix to negate (exclude)
• Example: music shibayan -bossa
  • ✅ Matches: /music/shibayan/album/song.mp3
  • ❌ Rejects: /music/shibayan/bossa-nova/track.flac

Name

• Search filename only (not full path)
• Example: demetori styx

Size

• Range: 10m-100m (10-100 MiB)
• Min: >5g or 5g-
• Max: <1k or -1k
• Suffixes: b, k, m, g

Date

• Range: 2023-01-01-2023-12-31
• After: >2023-06-01 or 2023-06-01-
• Before: <2023-01-01 or -2023-01-01

Tags

Search indexed metadata (requires -e2ts):

Tag Fields

• artist:, title:, album:
• genre:, date:, key:
• .bpm:, .dur:, .tn: (numeric)

Examples

artist: nhato
title: *dream*
.bpm: 128-140
genre: *trance*

Wildcards

• * matches any characters
• artist:*nhato* finds “nhato”, “Nhato”, “NHATO”

Raw SQL

Advanced queries using SQLite syntax:

Complex Queries

(tags like *nhato* or tags like *taishi*) 
and (not tags like *nhato* or not tags like *taishi*)

Finds songs by nhato OR taishi, excluding collabs.

substr(w,1,4) = 'abcd'

Search by partial hash.

Available Columns

• w - file hash (wark)
• sz - file size
• mt - modified time (unix timestamp)
• at - access time
• tags - concatenated metadata

Tag Search Examples

Find Specific Artist

artist: demetori

Find Songs in BPM Range

.bpm: 170-180

Find Long Songs

.dur: 600-

(Duration 600+ seconds = 10+ minutes)

Complex Tag Query

artist: *sound* -artist: *souledge*
genre: trance
.bpm: 138-142

Finds:

• Artist contains “sound” but not “souledge”
• Genre is trance
• BPM between 138-142

Raw SQL: Exclude Collabs

tags like *artist1* and tags like *artist2*
and (not tags like *artist1* or not tags like *artist2*)

Search Configuration

Control What’s Searchable

[global]
  # Exclude patterns from search results
  srch-excl: password|logs/[0-9]|backup

Dotfiles in Search

Hidden files (.filename) require special permission:

[/music]
  /mnt/music
  accs:
    r.: alice  # alice can see dotfiles
  flags:
    dotsrch    # dotfiles appear in search results

Without dotsrch, dotfiles are excluded from search even if user has . permission.

Per-Volume Search

Search behavior can differ per volume:

[/public]
  /srv/public
  flags:
    e2dsa      # index everything
    
[/private]
  /srv/private  
  flags:
    d2d        # disable all indexing
    
[/media]
  /srv/media
  flags:
    e2d        # index files
    d2t        # but not tags

Search Performance

Indexing Options

Full Index

copyparty -e2dsa -e2ts

• e2dsa: Index all files (writable + readonly)
• e2ts: Scan tags in all files without tags
• Best search performance
• Slow initial startup

On Upload Only

copyparty -e2d -e2t

• Only index new uploads
• Fast startup
• Won’t find existing files until uploaded again

Partial Index

copyparty -e2dsa --no-hash '\.iso$'

• Index metadata but skip hashing .iso files
• Faster scanning
• Cannot deduplicate or content-search skipped files

Exclude from Indexing

[/downloads]
  /mnt/downloads
  flags:
    noidx: '\.tmp$|\.part$'  # skip temporary files
    nohash: '\.iso$'          # hash-less indexing for ISOs

Database Location

Searches query the up2k.db SQLite database.

[global]
  hist: /fast/ssd/copyparty  # put db on SSD for speed
  
[/music]
  /slow/nas/music
  flags:
    hist: -                   # or per-volume: use volume's .hist/
    dbpath: /fast/ssd/music   # db only (thumbnails stay in volume)

Placing the database on an SSD dramatically improves search performance, especially for large libraries.

API Access

Search via HTTP API

# Search by name
curl 'http://localhost:3923/?q=demetori'

# Search with filters
curl 'http://localhost:3923/?q=styx&artist=demetori'

# JSON output
curl 'http://localhost:3923/?j&q=demetori'

Raw SQL via API

curl --data-urlencode "q=w like 'abc%'" \
  'http://localhost:3923/?raw'

Search Results

Results display:

• File path (click to navigate)
• File size
• Last modified date
• Metadata tags (if indexed)
• Upload time (if -e2d and -mte +.up_at)

Show Upload Time

[global]
  e2d
  mte: +.up_at  # append to existing tags

Or per-volume:

flags:
  e2d
  mte: +.up_at

Advanced: Periodic Rescans

Keep index current if other software modifies files:

[global]
  re-maxage: 3600  # rescan every hour
  
[/uploads]
  /mnt/uploads
  flags:
    scan: 300      # this volume: every 5 minutes

Rescans disable uploads temporarily. Delayed by --db-act (default 10s) during active uploads.

Filesystem Guards

Prevent indexing across filesystems:

[/music]
  /mnt/music
  flags:
    xdev   # don't cross filesystem boundaries
    xvol   # don't follow symlinks outside volume

• xdev: Skip bind-mounts and separate filesystems
• xvol: Skip symlinks outside volume top directory

Example: Music Library Search

[global]
  e2dsa                  # full file indexing
  e2ts                   # tag indexing
  mte: artist,title,album,.tn,.bpm,key
  mth: .dur,fmt,genre    # hidden but searchable
  hist: /fast/ssd/db     # fast database location
  
[/music]
  /mnt/nas/music
  accs:
    r: *
  flags:
    dotsrch              # include dotfiles in search
    scan: 3600           # hourly rescan
    srch_excl: 'temp|\._' # exclude temp and macOS files

Search examples:

artist: demetori .bpm: 170-180
title: *touhou* genre: *metal*
.dur: 300- artist: *sound* -artist: *collaboration*