OpenReader

The tool accepts EPUB, PDF, DOCX, Markdown and plain‑text files, processes them on the server, and provides a synchronized read‑along experience. As the text is spoken by a chosen text‑to‑speech engine, each word is highlighted in real time, and users can control playback speed independently of the synthesis model. The interface also allows exporting the spoken version as an MP3 or M4B audiobook with chapter information.

It targets users who want to listen to documents on any browser while keeping their reading progress, library organization, and preferences synced across devices. The application supports multiple OpenAI‑compatible TTS providers, including self‑hosted endpoints such as Kokoro, KittenTTS, and DeepInfra, giving flexibility in voice selection and cost control.

The project is a self‑hostable, open‑source service built with Next.js, offering Docker deployment, optional authentication, and configurable storage backends (embedded SeaweedFS or external S3‑compatible services). It is released under the MIT license and is considered stable.