pyLBRY

A Python library for reading from and archiving the LBRY network.

Requires Python 3.12 or later and a running lbrynet daemon. See docs/superpowers/specs/2026-08-02-pylbry-design.md for the design.

Install

uv add pylbry

Usage

from pathlib import Path

from pylbry import Archiver, Client

with Client() as client:
    channel = client.claims.resolve("lbry://@example")
    with Archiver(client, Path("archive")) as archiver:
        archiver.index_channel(channel.claim_id)
        archiver.fetch_pending(limit=10)

Or from the command line:

pylbry resolve lbry://@example
pylbry --root archive archive channel lbry://@example --limit 10
pylbry --root archive archive resume --limit 10
pylbry --root archive status

Archiving happens in two phases. Indexing records every claim in the channel into a SQLite database without downloading media, so the metadata survives even if the media never does. Fetching then downloads the queued claims.

Media downloads are large. Without --limit, archive channel fetches the entire channel, which is routinely tens of gigabytes.

Daemon setup

The default lbryum_servers list is mostly dead: LBRY Inc's spv*.lbry.com hosts refuse connections. If the daemon reports available_servers: 0 and its wallet never starts, pin it to a server that still runs, in ~/.local/share/lbry/lbrynet/daemon_settings.yml:

lbryum_servers:
  - hub.lbry.grin.io:50001

Entries must be host:port strings. The nested-list form that settings_get returns is silently discarded on load, yielding no servers at all.

Development

uv venv && uv sync
uv run pytest --cov-fail-under=100
uv run ruff check && uv run ty check

The suite has no mocks: every network path runs against a real daemon with its wallet started. Set PYLBRY_TEST_DAEMON if the daemon is not at http://localhost:5279, and PYLBRY_TEST_CHANNEL to draw test content from a channel other than the default. Tests reuse media the daemon already holds before starting any new download.