OpenReader review 2026: this open-source read-along document reader's TTS features, formats supported, self-hosting, pricing and alternatives.
OpenReader is an open-source, self-hosted read-along document reader created by independent developer Richard Roberson and first published on GitHub in January 2025. It is not a company or commercial product; it is a solo-maintained open-source project released under the MIT license, with a public demo hosted at openreader.richardr.dev.
The tool is designed for people who want a synchronized narration experience, audio read aloud with word-level highlighting, for their own documents, without depending on a subscription audiobook service.
OpenReader supports EPUB, PDF, DOCX, TXT, and Markdown files, using layout-aware parsing and ONNX Whisper-based alignment so that highlighting follows the true reading order even in complex, multi-column PDFs. Users can generate a full audiobook export from any supported document.
Text-to-speech can be powered by hosted providers such as OpenAI, Replicate, or DeepInfra, or by fully self-hosted, offline engines like Kokoro, KittenTTS, and Orpheus, giving users control over cost and data privacy. Deployment is via Docker, with support for S3-compatible storage and SQLite or Postgres databases.
OpenReader itself is completely free and open source under the MIT license; there is no subscription or license fee for the software. Costs, if any, come only from the text-to-speech provider a user chooses to connect, such as per-request API charges from OpenAI, Replicate, or DeepInfra.
Users who self-host an offline TTS engine like Kokoro can run OpenReader with no ongoing usage costs at all, aside from their own server hosting expenses.
Yes, OpenReader is free and open source under the MIT license. The only potential costs come from the text-to-speech provider you choose to connect.
OpenReader is built and maintained by independent developer Richard Roberson (GitHub handle richardr1126), first published in January 2025.
OpenReader supports EPUB, PDF, DOCX, TXT, and Markdown documents.
Yes, OpenReader can be configured with self-hosted, offline text-to-speech engines like Kokoro, KittenTTS, or Orpheus.
Yes, OpenReader can generate a full downloadable audiobook export from any supported document.
No, OpenAI is one optional text-to-speech provider; you can also use Replicate, DeepInfra, or a self-hosted TTS engine.
There is a public demo at openreader.richardr.dev, but there is no commercial SaaS version with a support agreement; the intended use is self-hosting.