Bindery

The Readarr replacement built to outlive its metadata sources.
Automated ebook & audiobook management for Usenet & torrents — monitor authors, search indexers, download, organize.

---

Mobile-friendly

       

---

Why Bindery?

You probably got here because Readarr died. The official project was archived in June 2025 when its metadata backend (api.bookinfo.club) went permanently offline — one dead endpoint took the whole app down. The community forks that followed lean on Goodreads scrapers that break whenever the site changes its HTML.

Bindery is built so that can't happen to you:

• No single point of failure for metadata. OpenLibrary leads, with Google Books, Hardcover, DNB, Audnex, and Audible behind it — all documented public APIs, zero scraping. If one source degrades, the others carry the library.
• Bring your dead Readarr install with you. Import readarr.db and your authors, indexers, download clients, and blocklist come across in one step. No re-entry, no CSV gymnastics.
• Boring to run, by design. A single Go binary with the UI embedded, SQLite, no external database, no sidecars, no nginx. Distroless image, Helm chart, and ARM builds down to a Pi Zero.

Clean-room Go rewrite, modern React UI, MIT-licensed, actively developed.

Tip

Coming from Readarr? Keep your library — point Bindery at the old database:

bindery migrate readarr /path/to/readarr.db

Authors are resolved again against your primary metadata provider and its fallbacks (Readarr's bookinfo.club backend is gone); indexers, download clients, and blocklist port across. Prefer the UI? Settings → Import. Full guide: Migrating from Readarr.

How Bindery compares

	Bindery	LazyLibrarian	Livrarr	Readarr
Status	Active (v1.37.x)	Active (community fork)	Alpha (0.1.x)	Archived (June 2025)
Stack	Go, single binary	Python	Rust	C# / .NET
Book metadata	6 independent sources, no scraping	Goodreads + fallbacks	4 sources, no scraping	Dead backend
Readarr .db import	Yes	No	No	—
Ebooks + audiobooks	Yes — separate roots, narrator metadata	Yes (+ magazines, comics)	Yes	Limited
ARM / Raspberry Pi	Yes, to Pi Zero (armv6)	Yes	arm64 only	Yes
Kubernetes / Helm	Yes, distroless	No	No	Community
Usenet + torrents	Both (SAB/NZBGet + qBit/Transmission/Deluge/rTorrent)	Both	Both (SAB + qBit)	Both

Pick Bindery for a modern, low-maintenance manager that survives metadata outages and swallows your old Readarr database whole. Pick LazyLibrarian if you want a long-established tool with magazine and comic support and don't mind a dated UI. Pick Livrarr if you want the same single-binary spirit and don't need the maturity, the Readarr import, Pi-class ARM, or a Helm chart yet. Calibre-Web / Audiobookshelf solve a different problem — serving an already-curated library, not automating acquisition — and pair well downstream of Bindery.

_{Competitor details verified June 2026 (Readarr archived 2025-06-27; LazyLibrarian active community fork; Livrarr v0.1.0-alpha5). Livrarr rechecked September 2026: v0.1.0-alpha6. Spotted something out of date? Open an issue.}

Features

Library management

• Author monitoring via OpenLibrary's author-works endpoint, configurable author monitor modes, all, future, latest or none as the install default plus a per-author mode that follows only the series you pick, per-book monitor toggles, and a wanted → imported workflow with live download progress on the Queue page.
• Unattended release discovery: monitored authors are checked for new books on a schedule, so a followed author's next book joins the library without a click. It ships off; turn it on in Settings, General, New release discovery by picking Daily, Weekly or Monthly. Opt an author out with Monitor new items: Don't add them.
• Dual-format books — each title holds an ebook and an audiobook in independent slots, with separate search, grab, and import pipelines, and the audiobook side moves multi-part .m4b / .mp3 folders as one unit.
• Series support with position tracking and search by series title, edition tracking (format / ISBN / publisher / page count), Calendar view of upcoming releases, and multiple library roots.
• Library scan matching in five passes: exact ASIN, then a backwards Author - Title filename whose folder author the catalogue already knows, then the filename's title, then the book folder's title, and finally series name plus position. Every title pass is one fuzzy comparison that has to agree on the author as well. Honours librarian sort-suffix form (Title, The) and series-annotated filenames ([Mistborn, Book 1]).
• Library adoption: whatever the scan could not match waits on the Import page as one row per book (an audiobook folder or disc set is one row), with suggested matches (one click Confirm only for a strong match), one "Add author" decision per missing author, and Undo. Files are registered where they are, never moved.
• Author aliases (RR Haywood / R.R. Haywood / R R Haywood merge into one canonical row), and metadata re-bind to correct a wrong match without delete-and-re-add. On a book’s File → Re-bind dialog, search configured metadata providers and select an OpenLibrary work or Hardcover book, or enter an exact provider ID and press Enter. Results already in the library are marked and link to the existing book; records belonging to another library book cannot be selected. Each result’s Links menu opens its upstream page when available; OpenLibrary editions can be viewed but require a work ID for re-binding. Re-binding updates metadata and series without moving files; author mismatches require confirmation.
• Explicit author-catalogue reconciliation with a selectable preview: remove chosen stale metadata-only Wanted rows after changing provider or metadata profile while always protecting imported books and every row with a tracked file.
• Manual metadata editing with field locks: edit a book's title, description, genres, language or release date and the edit survives every refresh. Fix match reassigns a file that was attached to the wrong book. See docs/Metadata-Editing-Wiki.md.

Search & downloads

• Newznab + Torznab indexers queried in parallel, deduplicated, then composite-ranked by format quality, edition tags (RETAIL / UNABRIDGED / ABRIDGED), year match, grab count, size, and ISBN exact-match bonus.
• Smart matching — four-tier query fallback (t=book → surname+title → author+title → title), word-boundary keyword matching, ordered title words with numeric qualifiers preserved, rejection of conflicting trailing author credits, dual-author-anchor for ambiguous short titles, subtitle-aware (Title: Subtitle). Title-only releases remain supported.
• SABnzbd, NZBGet, qBittorrent, Transmission, Deluge, rTorrent/ruTorrent — with Use SSL and URL Base for reverse-proxy subpaths.
• Diagnose on each saved download client: one click checks the connection, the category, where ebook and audiobook grabs actually land, the path remap, whether Bindery can read that folder, and whether imports can hardlink into each library folder, then names the first thing to fix. No shell needed, which matters in the distroless image.
• Auto-grab sweep on a configurable interval, 12h by default, immediate search on add or wanted flip, plus interactive per-book search and "Search all wanted" per author. Global kill-switch pauses auto-grab without losing your monitored list.
• Quality profiles covering every format release parsing recognises (EPUB, MOBI, AZW3, PDF, plus AZW, DJVU, CBR, CBZ, FB2, LIT, RTF, TXT and the audio containers M4B, M4A, FLAC, MP3, OGG), language filter, regex-based custom formats, delay profiles, blocklist (consulted on every search; one-click add from History), and failure visibility in Queue and History.
• Indexer-friendly by default: an optional per-indexer Daily query limit with usage shown on the Indexers tab, a rate-limited indexer is left alone for longer each time it refuses (one hour, then three, six, twelve and a day) and its row says when searches resume, an indexer that has stopped working is flagged in Settings, and a per-indexer freeleech only policy holds ratio-costing releases for manual approval.

Import & organize

• Completed downloads matched by NZO ID and placed in the library with configurable naming. Modes: Auto (default — hardlink when possible, else copy, so a torrent keeps seeding; a usenet grab is moved, since there is nothing to seed), Move, Copy (keep source for seeding), Hardlink (zero extra disk; same filesystem required), External (hand off to a sibling tool).
• Naming tokens — {Author}, {SortAuthor}, {Title}, {Year}, {Series}, {SeriesNumber}, {Genre}, {Lang}, {ASIN}, {ext}, plus {Part} in the audiobook template — collapse cleanly for non-series books, with conditional literals ({Title}{ - Series} emits the dash only when a series exists) and zero-pad widths ({SeriesNumber:2} → 02).
• Cross-filesystem-safe moves: atomic rename when possible, copy + verify + delete for NFS / separate volumes. Full grab / import / failure history per book.
• The Import page is where files you already have come in, two ways: In your library is the adoption list described above, and From a folder points at a folder anywhere Bindery can read, matches what it finds and imports it into the library, creating a book from a metadata search when nothing in the catalogue fits.
• Calibre integration: three ways to hand a book over, plus a read side. Register each import with Calibre through the calibredb CLI or the Bindery Bridge plugin (cross-container), mirror each ebook into a Calibre-Web-Automated ingest folder, or hand off through a drop folder so CWA or Calibre owns the library. Separately, read an existing Calibre library's metadata.db as Bindery's catalogue. See docs/Calibre-Integration-Wiki.md.
• Audiobookshelf import — pull an existing ABS server's book libraries in as Bindery's catalogue (metadata-first, dry-run, review queue for ambiguous matches, rollback), with an ABS library-scan trigger after every audiobook import. See docs/ABS-Import-Wiki.md.
• Grimmory push (preview) — imported ebooks are sent to a self-hosted Grimmory library via its BookDrop inbox, with a bulk Push all for existing files.
• metadata.opf sidecar (opt-in) — write a Calibre-style metadata.opf next to each imported book, carrying Bindery's own canonical title/author/series/identifiers/etc. so a library app that reads sidecar metadata sees consistent data regardless of which source the file came from. Refreshed on Reorganize.

Metadata sources — all stable, documented, public APIs. No Goodreads scraping.

Source	Auth	Used for
OpenLibrary	None	Default primary: authors, books, editions, covers, ISBN
Google Books	API key (free)	Enrichment: descriptions, ratings
Hardcover.app	API token (free)	Search enrichment + community ratings, series, wishlist; can be promoted to primary for a curated, translation-free catalogue. Token required for all queries — without it Hardcover is silently skipped (troubleshooting)
DNB	None (public SRU)	German-language descriptions, language, year, publisher; can be promoted to primary
Audnex	None	Audiobook narrator, duration, cover by ASIN
Audible	None	Supplemental audiobook author lookup — pulls ASINs OL/Hardcover miss

Cover images are fetched and cached server-side under <dataDir>/image-cache/ (30-day TTL). Every imageURL is rewritten to /api/v1/images?url=... before leaving the server — the browser never contacts third-party image hosts directly.

Discover — personalised recommendations on the Discover page from multiple signals: next-in-series for what you're reading, new releases from monitored authors, genre similarity (≥ 20 books in library), OpenLibrary subject popular picks, and Hardcover wishlist cross-reference. Recency scoring is relative to the median publication year of your library, so backlist readers aren't penalised. Hard-filters owned, dismissed, excluded-author, wrong-language, sub-3.0-rated and omnibus titles, and drops anything under 50 ratings when nothing else vouches for it. Dismiss / exclude actions persist.

Migration — upload readarr.db directly (authors resolved again against your primary metadata provider and its fallbacks since bookinfo.club is dead; indexers, download clients, and blocklist port structurally), or paste a newline-separated list of author names. CLI: bindery migrate {csv,readarr} <path> for first-time bulk imports without opening the UI.

Operations

• Authentication — first-run setup creates an admin account (argon2id, signed session cookies). Four modes: Enabled / Local only (bypass for private IPs) / Disabled / Proxy (trust upstream X-Forwarded-User from a configured trusted proxy — drop-in for Authelia / Authentik / oauth2-proxy). Per-account API key, per-IP login rate limiting, CSRF double-submit (API-key clients exempt).
• OIDC — native Authorization Code + PKCE with multi-provider support. Pre-configured for Google, GitHub (via Dex), Authelia, and Keycloak; identifies users by stable (issuer, sub) so email/username changes don't break logins.
• Multi-user mode — per-user libraries, monitored authors, profiles, and downloads. Admin role manages indexers / download clients / users; standard users see only their own catalogue. Local, OIDC-provisioned, or forward-auth-mapped.
• Requests (requester role). Give family and friends an account that can browse the library read only and ask for a book or an author, but cannot grab, download, delete or configure anything. Admins approve from a Requests queue with the usual add choices, and the added books belong to the person who asked. Requesters can follow each request to "available", and a webhook can announce new ones. See docs/multi-user.md.
• Webhook notifications for grab, import, failure, new books found, new requests, health changes and a new Bindery release (pipe to Apprise, ntfy, Home Assistant, Discord, Slack via proxies). On-demand SQLite backups. Persistent log viewer in Settings → Logs with runtime DEBUG toggle.
• Download logs from Settings → Logs as a text file for bug reports. In-app update badge when a newer release exists. bindery db-check and bindery db-repair report and repair orphaned database rows offline, for an instance that will not start (details).
• Arr-compatible queue at GET /api/queue for Harpoon and other *arr-aware tools — pagination, sort, live size, status, client, remote ID, protocol.

UI

• Modern React 19 + TypeScript + Tailwind CSS SPA with search-first author acquisition and deep-linkable routed /book/:id and /author/:id pages.
• Five entries in the top bar: Library (Authors, Books, Series), Activity (Wanted, Queue, History, and Requests for an admin), Import, Calendar and Discover. A group opens on its first page and repeats the rest as tabs above the content, and every page keeps the address it always had.
• Library search in the header over your authors, books and series, with a last row that hands a miss to a single Add to library dialog for authors and books (title, author, ISBN or ASIN).
• Light / dark themes (respecting prefers-color-scheme first paint), grid / table view toggles, mobile-friendly responsive layout, hamburger nav, agenda-style mobile Calendar.
• Full pagination, search, filter, and sort on every list page; preferences persist to localStorage.
• Previous/Next navigation between authors on the author detail page, stepping through the list page you came from.
• Previous/Next navigation between books, stepping through the list page you came from (Books, an author's own book list, or Wanted).
• 8 languages — English, French, German, Dutch, Spanish, Filipino (Tagalog), Indonesian, Korean — auto-detected from the browser, override in Settings.
• OPDS 1.2 catalogue at /opds/ for KOReader, Moon+ Reader, and other reading apps. Sign in with your Bindery username and password over HTTP Basic, or pass your API key as an X-Api-Key header or an ?apikey= query parameter.

Packaging — single Go binary with the React frontend embedded via go:embed. Distroless container, non-root by default, and a Helm chart that adds a read-only root filesystem, all capabilities dropped and a RuntimeDefault seccomp profile. Helm chart for ArgoCD / Flux. Pure-Go SQLite via modernc.org/sqlite — no CGO, no external database.

Quick Start

Docker

docker run -d \
  --name bindery \
  -p 8787:8787 \
  -v /path/to/config:/config \
  -v /path/to/books:/books \
  -v /path/to/downloads:/downloads \
  ghcr.io/vavallee/bindery:latest

Open http://localhost:8787, follow the first-run setup to create the admin account, and you're in.

Docker Compose

services:
  bindery:
    image: ghcr.io/vavallee/bindery:latest
    container_name: bindery
    user: "1000:1000"   # the uid:gid that owns your books and downloads
    ports:
      - 8787:8787
    volumes:
      - ./config:/config
      - /media/books:/books
      - /media/downloads:/downloads
    environment:
      - BINDERY_LOG_LEVEL=info
    restart: unless-stopped

The image is distroless and cannot switch users at runtime, so user: is how you set the UID that writes your library. BINDERY_PUID and BINDERY_PGID only assert that it came out right.

Other install methods

Pre-built binaries for Linux (amd64, arm64, armv7, armv6), macOS and Windows (amd64, arm64), Kubernetes (Helm chart at charts/bindery/), and the Unraid Community Applications template are all covered in docs/DEPLOYMENT.md — including UID/GID setup, path remapping for multi-container deployments, the full environment-variable reference, and per-version upgrade notes.

Configuration

Bindery is configured through the web UI under Settings — indexers, download clients, quality profiles, naming, notifications, auth, and everything else runtime-tunable lives there. A small set of bootstrap-only knobs are environment variables:

Variable	Default	Purpose
BINDERY_PORT	8787	HTTP server port
BINDERY_DB_PATH	platform-default	SQLite database path
BINDERY_DATA_DIR	platform-default	Config directory (backups, image cache, secrets)
BINDERY_LIBRARY_DIR	/books	Imported ebook destination
BINDERY_AUDIOBOOK_DIR	inherits library	Imported audiobook destination
BINDERY_DOWNLOAD_DIR	/downloads	Where the download client deposits completed jobs
BINDERY_AUDIOBOOK_DOWNLOAD_DIR	inherits download dir	Separate watch folder for audiobook downloads
BINDERY_URL_BASE	(empty)	Reverse-proxy subpath (e.g. /bindery)
BINDERY_OUTBOUND_PROXY	(empty)	Route outbound HTTP (indexers, metadata, covers, notifications, telemetry) through an http/https/socks5 proxy. LAN/loopback destinations bypass it by default — see DEPLOYMENT.md
BINDERY_LOG_LEVEL	info	Log verbosity: debug, info, warn or error
BINDERY_PUID / BINDERY_PGID	(unset)	Sanity-check assertions for the container UID/GID. They only assert; the distroless image does no runtime user switching, so set the UID with --user or Compose's user: as well

The full reference (path remapping, API-key seeding, telemetry, trusted-proxy, rate-limit knobs, cookie-Secure policy) is in docs/DEPLOYMENT.md#environment-variables. OIDC and forward-auth proxy variables live in docs/auth-oidc.md and docs/auth-proxy.md.

Supported integrations

Category	Implementations
Usenet clients	SABnzbd, NZBGet
Torrent clients	qBittorrent, Transmission, Deluge, rTorrent / ruTorrent (XML-RPC over HTTP or SCGI)
Indexers	Newznab (NZBGeek, NZBFinder, NZBPlanet, DrunkenSlug, …), Torznab (Prowlarr, Jackett, direct endpoints), with per-indexer category overrides
Metadata sources	OpenLibrary, Google Books, Hardcover, DNB, Audnex, Audible
Import lists	Hardcover lists, synced on a configurable interval
Library tools	Calibre (calibredb, Bindery Bridge plugin, metadata.db import), Audiobookshelf (import and scan trigger), Grimmory (preview), post-import drop folder for Calibre-Web-Automated and similar (guide)
Notifications	Generic webhooks — pipe to Apprise / ntfy / Home Assistant / Slack / Discord
Authentication	Local (argon2id), API key, OIDC (Google, GitHub via Dex, Authelia, Keycloak, …), forward-auth proxy
Reading apps	OPDS 1.2 catalogue at /opds/ (KOReader, Moon+ Reader, Aldiko, …)

All download clients support Use SSL and URL Base for connections through a reverse-proxy subpath.

Architecture

Bindery is a single Go binary (chi router, distroless container) with the React 19 + TypeScript frontend embedded via go:embed, talking to SQLite in WAL mode through the pure-Go modernc.org/sqlite driver — no CGO, no external database, no sidecars.

   Newznab / Torznab
      indexers
         │
         ▼
┌────────────────────────────┐
│         Bindery            │──► SABnzbd / NZBGet / qBit / Transmission / Deluge / rTorrent
│  Go backend + React SPA    │──► /books/ library
│  SQLite (WAL mode)         │──► Webhook notifications
└────────────────────────────┘
    ▲                    ▲
    │                    │
OpenLibrary      Google Books, Hardcover.app, DNB, Audnex, Audible
(default primary)          (enrichers)

Component breakdown, package layout, concurrency model, and design rationale are in docs/ARCHITECTURE.md.

API

Every feature is exposed under /api/v1/*, with an arr-compatible /api/queue for external tools and an OPDS catalogue at /opds/. Quick taste:

# Add an author by OpenLibrary ID and start monitoring
curl -X POST -H "X-Api-Key: $KEY" -H "Content-Type: application/json" \
  -d '{"foreignAuthorId":"OL23919A","monitored":true,"searchOnAdd":true}' \
  http://bindery:8787/api/v1/author

# List wanted books
curl -H "X-Api-Key: $KEY" "http://bindery:8787/api/v1/book?status=wanted"

The full endpoint catalogue, authentication rules (API key, session cookie, local-only, OIDC, proxy), CSRF semantics, and integration examples are in docs/API.md.

Documentation

Topic	Where
Quickstart — zero to first download, end to end	docs/QUICKSTART.md
Quickstart (wiki) — first author to first grab in 10 minutes	Wiki
User guide — how the app thinks: catalogue-first flow, wanted vs monitored, imports	docs/User-Guide-Wiki.md
Deployment — Docker, Compose, k8s/Helm, binary, UID/GID, env vars, upgrades	docs/DEPLOYMENT.md
Architecture — components, data flow, dependencies	docs/ARCHITECTURE.md
API — REST endpoints, auth, integration patterns	docs/API.md
Search and matching design (contributors): how titles and author names are normalised and compared	docs/search-design.md
Roadmap — planned work and explicitly-out-of-scope items	docs/ROADMAP.md
Multi-user — roles, user management, CSRF tokens	docs/multi-user.md
Upgrading to v1.0 — multi-user migration: backup, dry-run, rollback	docs/upgrade-v1.md
OIDC auth — Google / GitHub / Authelia / Keycloak setup	docs/auth-oidc.md
Forward-auth (proxy) auth — Authelia / Authentik / oauth2-proxy	docs/auth-proxy.md
Auth troubleshooting — symptom → cause → fix across all auth modes	docs/troubleshooting-auth.md
ABS import (overview) — what gets imported, setup, review queue, rollback	docs/ABS-Import-Wiki.md
ABS import (reference) — implementation detail, mapping rules, API surface	docs/abs_import.md
Enhanced Hardcover series — token setup, series linking, catalog diffs, missing-book fill	docs/Hardcover-Series-Wiki.md
Calibre and CWA: the three hand off topologies, what each does on disk, troubleshooting	docs/Calibre-Integration-Wiki.md
Storage & hardlinks — single-mount layout, import modes, per-author audiobook root	docs/Storage-And-Hardlinks-Wiki.md
Migrating from Readarr — readarr.db import, Goodreads CSV import, library scan	docs/Migrating-From-Readarr-Wiki.md
Metadata editing — manual edits, field locking, bulk genre overrides	docs/Metadata-Editing-Wiki.md
Troubleshooting — permission-denied, path-remap, import failures	docs/Troubleshooting-Wiki.md
Contributing & CI checks — dev setup, full quality/security matrix, local check suite	CONTRIBUTING.md
Changelog — release notes	CHANGELOG.md
Reverse-proxy & SSO setups — Traefik / Caddy / Nginx / Authelia / Authentik recipes	Wiki
Indexer & download-client recipes — NZBGeek / DrunkenSlug / Prowlarr / Jackett / SAB / qBit tips	Wiki

Community

• Discord — real-time help, setup questions, release chat: discord.gg/RpuYYRM9cZ. The #support channel is the best place to ask; #releases is updated on every release. The three read-only voice channels at the top of the server show live active-install count, latest release, and GitHub star count, refreshed from the telemetry API every 10 minutes.
• GitHub Issues — bug reports and feature requests: issues.
• GitHub Discussions — open-ended design questions, show-and-tell, integration recipes: discussions.

Please keep security reports out of Discord and public issues — see SECURITY.md for the private disclosure channel.

Security

Bindery holds API keys, reaches LAN services, and writes to disk — we take that seriously. Every push and weekly cron runs govulncheck, CodeQL, Semgrep, gitleaks, Hadolint, Checkov, Grype, Syft, ZAP baseline, and OpenSSF Scorecard, with gosec enforced inside the blocking golangci-lint gate, and findings published as SARIF in GitHub's Security tab. Release images ship with SLSA build provenance and Syft SBOMs. In-app: SSRF guards on every outbound URL, hardened response headers (CSP, frame-deny, referrer-policy, auto-HSTS), distroless non-root read-only rootfs container with all caps dropped, and digest-pinned base images.

To report a vulnerability, follow the process in SECURITY.md. The full threat model and verification recipes live on the wiki Security page.

Telemetry

Bindery sends one anonymous ping per day to api.getbindery.dev so the maintainer can count active installs and spot widespread breakage. The payload contains a random install_id (generated on first run), the binary version, os, arch, and deploy method (kubernetes/docker/binary); a features section of counts and booleans describing which subsystems are configured (never names, URLs, or values); and an errors section with the number of ERROR/WARN log entries over the last 24 hours plus the five most frequent error messages. The error messages are the fixed, developer-written log message strings only (truncated to 120 characters) — log details (attrs) such as titles, paths, URLs, or usernames are never sent. No hostnames, IP addresses, library contents, or personal data. The response carries the latest published version, which powers the in-app update notice (an amber dot on the account icon in the header, the version inside its menu, and Settings → About). Opt out with telemetry.enabled: false in Settings → Logs, or BINDERY_TELEMETRY_DISABLED=true before first run — either switch disables the entire ping, including the error counters and the update badge (a disabled ping means the app has no way to learn a newer version exists).

Retention, the legal basis, how IP addresses are handled, and how to have an install row deleted are in PRIVACY.md.

Contributing

PRs, issues, and feedback welcome. See CONTRIBUTING.md for the dev setup, the full local check suite, and the PR flow. Tracked feature work lives in docs/ROADMAP.md — open an issue before starting anything substantial.

License

MIT. See LICENSE for details.

Bindery statically links its Go dependencies and embeds the compiled web UI, so the third-party attribution ships with every artifact: THIRD_PARTY_LICENSES.md lists every module and npm package in the released binary with its license, reproduces the NOTICE files the Apache-2.0 dependencies carry, and is included in the release archives and at /THIRD_PARTY_LICENSES.md inside the container image. The file is generated — make licenses regenerates it and CI fails the build when it drifts from the dependency set, so run it in the same PR whenever you add, drop, or bump a dependency.

Dependency licences govern what the binary may link. What a deployment may do with the data Bindery fetches at runtime is a separate question, and docs/third-party-data.md records it per source — including that a commercial deployment has to exclude the aggregated Hardcover ratings. Self-hosting for yourself is unaffected.

Disclaimer

Bindery is not affiliated with or endorsed by Readarr, Audible, Amazon, Goodreads, Calibre, or Prowlarr. All trademarks are the property of their respective owners.

Bindery is a download manager. It finds and organises files from the indexers and download clients you configure, and it ships with none. You are responsible for the legality of what you acquire with it and for complying with the terms of the services you point it at.

Acknowledgments

• The *arr community for pioneering the monitor-search-download-import pattern
• OpenLibrary for free, open book metadata
• The Readarr project for the original vision, even though the implementation couldn't be sustained