Documentation
Read the copied guides, API reference and update-service notes.
ReadAIrr keeps the familiar Readarr automation base while moving toward narrator-aware matching, STT-backed review, hosted metadata choices, Docker/Postgres deployment, and a public update path.
ghcr.io/readairr/app, with a same-VM Postgres compose path and Unraid packaging work underway.
ReadAIrr adds visible narrator navigation and is building a missing-narrator review flow that can use metadata and speech-to-text evidence together. The goal is to let users accept high-confidence matches quickly while keeping manual confirmation close to the transcript when judgment matters.
Unmatched files are being shaped around real audiobook folders: multi-part books can be treated as one review entry, numbered files can be grouped together, and STT only needs to sample the first file in a series before proposing evidence.
New docker deployments default to the hosted Goodreads-compatible metadata service. Development settings also offer a hosted Hardcover source and a custom rreading-glasses URL for people who want to run their own metadata service.
Public dev, val, and prod container tracks pair with a first-party update service, a Docker Compose path with Postgres included, and an Unraid template that exposes the options users actually need.
The latest development work gives the project public Docker tracks, hosted metadata defaults, cleaner deployment docs, narrator-focused workflows, and a site that can explain the direction clearly.
dev, val, and prod images pair with Compose and Unraid setup paths users can reason about.
ReadAIrr is distributed through public GitHub Container Registry images. Choose the track that matches how much change you want to absorb.
View Docker image tracksghcr.io/readairr/app:prod for normal installs and public testing.
ghcr.io/readairr/app:val for release-candidate checks before production promotion.
ghcr.io/readairr/app:dev for active development builds and newest feature testing.