# absdy — audiobookshelf provider Mounts a self-hosted [audiobookshelf](https://www.audiobookshelf.org/) (ABS) server at `/abs`, so you can browse, search, and play your audiobooks from crabidy. See [the design doc](../architecture/audiobookshelf-provider.md) for the design and decisions. ## Configuration — `~/.config/crabidy/abs.toml` The provider needs your server URL and an API key. Without both, `/abs` does not mount (the rest of the server is unaffected). ```toml # Required. base_url = "https://audiobookshelf.example.com" api_key = "" # Optional (defaults shown). items_per_library = 200 # books listed per library search_results = 50 # books listed per search term call_timeout_secs = 30 # per-request timeout ``` Create an API key in audiobookshelf under **Settings → Users → (your user) → API Keys** (or **Settings → API Keys** on older versions). The key is a secret: it is redacted from logs and never printed, but keep `abs.toml` private anyway. ## The library tree ```text /abs └── one node per "book" library ├── search create a search term with `%` │ └── books matching the term │ └── the book's audio files as tracks └── an audiobook; its files are tracks └── a track (one audio file) ``` - A **library** lists its books plus a creatable `search` node. Search is per-library — a term created under one library does not appear under another. - A **book** lists its audio files as tracks; queue or capture (`W`) the whole book, or a single file. An ebook-only item (no audio) is shown but is not queueable. - Podcast libraries are not shown (audiobooks only, for now). ## Notes - Playback streams each file directly from ABS with a per-file token URL and HTTP range requests — no transcoding session. - Listings are capped (see the config) and fetched fresh; only your typed search terms are remembered, in memory, until the server restarts.