bandcamp-explorer

A terminal browser and Python library for Bandcamp.

Search for artists and albums, discover releases by genre and location, browse artist/label profiles and discographies, all from the command line.

Install

Requires Python 3.12+.

Terminal CLI

uv tool install bandcamp-explorer
# or
pip install bandcamp-explorer

Discord bot

uv tool install bandcamp-explorer[discord]
# or
pip install bandcamp-explorer[discord]

Create a bot application at the Discord Developer Portal, enable the bot scope with Send Messages and Use Slash Commands permissions, then invite it to your server with the generated OAuth2 URL.

Set your bot token and run:

export DISCORD_TOKEN=your-bot-token
bandcamp-discord
# or with a .env file in the current directory
bandcamp-discord

Use --guild GUILD_ID to sync slash commands instantly to a specific server (global sync can take up to an hour).

Slash commands (all under /bandcamp):

Development

git clone https://github.com/gabriel-jung/bandcamp-explorer.git
cd bandcamp-explorer
uv sync

CLI

Search

bandcamp "caladan brood"                  # search everything
bandcamp "erang" --artist                 # artists/labels only
bandcamp "echoes of battle" --album       # albums only
bandcamp "a forest whisper" --track       # tracks only

Browse by tag

bandcamp --tag dungeon-synth                        # newest arrivals (default)
bandcamp --tag black-metal --top                    # best-selling
bandcamp --tag dungeon-synth --rand                 # surprise me
bandcamp --tag dungeon-synth --location france
bandcamp --tag dungeon-synth --location paris
bandcamp --tag dungeon-synth black-metal            # multi-tag

Slices: --new (default), --top, --rand.

Locations are resolved to geoname IDs via Bandcamp's autocomplete and cached locally; force a refresh with --refresh-location.

Direct URLs

bandcamp https://erang.bandcamp.com/album/tome-iv
bandcamp https://erang.bandcamp.com

Interactive navigation

After selecting a result, you enter an interactive browser:

• Artists: view bio, browse discography, select an album to see its tracklist, select a track to view its page, navigate to the label.
• Albums: header with tracklist, description, and lyrics; navigate to the artist/host page or select a track.

Press 0 to go back, Ctrl+C to quit.

Output modes

bandcamp "erang" --artist --json            # output as JSON
bandcamp "erang" --limit 10                 # cap results
bandcamp --tag dungeon-synth --json --limit 100   # cap tag dump
bandcamp https://erang.bandcamp.com/album/tome-iv --json
bandcamp https://erang.bandcamp.com/album/tome-iv --full   # all sections at once
bandcamp -v ...                             # enable debug logging

Terminal images

Album covers and artist images render inline on terminals that support the iTerm2 or Kitty image protocol (iTerm2, Kitty, WezTerm, Mintty).

Library

The core module has no terminal dependencies; use it in scripts, pipelines, or other tools. All data is returned as plain dicts with a _type discriminator key.

from bandcamp_explorer.core import (
    BandcampClient, AlbumAPI, ArtistAPI, DiscoverWebAPI, SearchAPI,
    NotFoundError, resolve_geoname,
)

with BandcampClient() as client:
    # Search (one call returns the whole result set)
    results = SearchAPI(client).search("caladan brood", item_type="album")

    # Discover releases by tag (new discover_web endpoint)
    discover = DiscoverWebAPI(client)
    releases, cursor, total = discover.discover(tags=["dungeon-synth"], slice_="new")
    all_releases = discover.discover_all(tags=["dungeon-synth"], max_pages=3)

    # Fetch album details (skip cover-art bytes with fetch_art=False)
    album = AlbumAPI(client).get("https://erang.bandcamp.com/album/tome-iv")
    for track in album["tracks"]:
        print(f"  {track['position']}. {track['title']} ({track['duration']})")

    # Fetch artist/label profile
    artist = ArtistAPI(client).get("https://erang.bandcamp.com")
    for item in artist["discography"]:
        print(f"  {item['title']}")

    # Location filtering (geoname-based)
    geoname_id = resolve_geoname(client, "paris")
    releases, _, _ = discover.discover(tags=["dungeon-synth"], geoname_id=geoname_id)

    # Download images
    client.download_image(album.get("image_url"), output_dir="./images/")

Errors

AlbumAPI.get and ArtistAPI.get raise NotFoundError when a page 404s, so callers can tell a deleted release from a failed fetch. Every other transport failure returns None. If Bandcamp answers with its bot-defence interstitial (HTTP 200 with no content in it), the client raises ChallengeError and then fails fast for two minutes rather than hammering a blocked endpoint. Never treat a ChallengeError as a missing resource; it means "ask again later".

from bandcamp_explorer.core import ChallengeError, NotFoundError

try:
    album = AlbumAPI(client).get(url)
except NotFoundError:
    ...  # gone for good, stop retrying
except ChallengeError:
    ...  # blocked for now, retry later

TLS fingerprints

Bandcamp soft-blocks some TLS fingerprints by answering HTTP 404 to pages a different fingerprint fetches fine, so a bare 404 is not proof of deletion. The client re-checks every 404 against a short list of known-good fingerprints before raising NotFoundError. If one of them serves the page, that fingerprint takes over the session for the rest of the client's life, so the extra request is paid once rather than on every later 404.

# Pick the fingerprint yourself (default: curl_cffi's floating "chrome" alias).
client = BandcampClient(impersonate="chrome124")

# Change or disable the re-check ladder.
client = BandcampClient(fallback_impersonate=("chrome131", "chrome124"))
client = BandcampClient(fallback_impersonate=())  # every 404 raises at once

client.impersonate  # the fingerprint currently in use, after any promotion

Names come from curl_cffi's impersonate targets; entries the installed version does not know are skipped rather than raising. Which fingerprints are blocked varies by vantage point: two hosts can see the same cutoff between builds with the sides swapped, one serving the recent ones and challenging the old, the other the reverse. Measure from the machine that will run the fetches with python scripts/probe_fingerprints.py, which reads response bodies rather than status codes because a blocked fingerprint answers HTTP 200 with the interstitial.

The ladder only ever refuses to believe a 404, it never invents one. A fallback that errors, is challenged, or is unknown proves nothing, so it is skipped and the original 404 stands. A challenged fallback never arms the client's challenge backoff either: that session is a throwaway and says nothing about the primary.

NotFoundError.confirmed_by names the fallback fingerprints that independently saw the same 404. It is empty when nobody could check, which is much weaker evidence than a 404 two working fingerprints agreed on. Callers that flag rows deleted should require it:

try:
    album = AlbumAPI(client).get(url)
except NotFoundError as e:
    if e.confirmed_by:
        mark_deleted(url)  # two independent fingerprints agree it is gone
    else:
        ...  # nothing could corroborate it; leave the row alone and retry

The default ladder deliberately spans browser families. Two Chrome builds share a failure axis: Bandcamp splits Chrome between 131 and 133a, and which side is served depends on where you fetch from, so a pair of Chrome fallbacks can land on the blocked side together and rescue nothing. Firefox and Safari are off that axis.

Bandcamp removed the dig_deeper hub endpoint, so DiscoverAPI was dropped in 0.6.0; use DiscoverWebAPI. resolve_location went with it: it resolved Bandcamp's internal location tag ids, which only that endpoint accepted. DiscoverWebAPI filters by geoname_id, so resolve_geoname is the one you want.

License

MIT