System TTS · free
Your phone's built-in speech engine, used directly: no key, no cost, and offline once an offline voice pack is installed. For quiet narration it is already listenable — build the habit on the free tier first.
AI listening
Your eyes are on the subway rail.
Let your ears keep the book.
Any imported txt or epub can be listened to: system TTS for free (offline with an offline voice pack installed), or online AI voices through your own key. Chunked chapter cache, 0.5–2.5x speed, sleep timer — and with the reader open, the page follows the narration and the position is saved as you go.
- System TTS: free
- Online voices: BYOK, direct to provider
- Reader follows the narration
Play/pause · change speed · jump chapters · open the engine / voice / timer / cache sheets
Listening panel · live prototype
9:41
Start free, upgrade only if your ears ask for it
Online AI voices · BYOK
Connect your own OpenAI-compatible speech endpoint and type any voice name (nova, alloy, or a provider-specific one). Costs go on your key at provider list price (rates depend on provider and text length), and every synthesis call is itemized in the usage log.
Change the voice, keep the book
A young neutral voice for fast arcs, a deeper one for historical interludes. Voice and speed are per-book settings; after a voice change the chapter is re-synthesized in the new voice.
Subways have no reliable network, so caching is a first-class feature
Online voices synthesize in chunks: while one chunk plays, the next ones queue in the background — the panel shows a live “3/12 chunks cached” strip. On Wi-Fi, tap “download the chapter” and carry tomorrow's commute in your pocket: zero-data playback, tunnels included.
The cache also guards your wallet: synthesized audio is stored keyed by chapter + voice + model, so as long as the cache is present, replays never re-synthesize. Only an actual voice change re-synthesizes; clearing the cache or switching devices means the affected chapters synthesize once more.
With the book open, the page follows the voice
Start narration inside the reader and the paragraph being read scrolls into view — in page mode, the pager turns to that page. Close your eyes for a stretch; when you open them the page is where your ears got to, and that position is saved into your reading progress as it moves. From the listening screen, one tap on “Text” drops you back into the reader.
The commute staples are all here: sleep timer (15/30/60/90 minutes), lock-screen controls, and continuous 0.5–2.5x speed.
Where AI narration honestly stands
Single-narrator prose is close enough to a human audiobook that most people stop noticing after a chapter. The gap is multi-character dialogue: human narrators switch voices per character; AI narration is one voice throughout. Fast-paced webnovels feel fine; ensemble drama will make you miss a human cast.
Wild txt files are full of things that should never be spoken — scene dividers, vote-for-me footers, emoji. We strip those before synthesis, so you will not hear a warm voice solemnly pronounce three asterisks. Rare names still trip occasionally; ears adjust within a chapter or two.
The full three-week field test: turning an epub into an AI audiobook
FAQ
Common questions
Does listening cost money?
System TTS is completely free (the phone's built-in engine; offline with an offline voice pack installed). Online AI voices go on your own API key at provider list price — most providers bill per synthesized character — and chapters replay from cache without re-synthesizing. On the BYOK path Foreverse adds no markup and requests go directly to the provider.
Which books can be listened to?
Any txt or epub you import. TXT chapters are detected automatically, epub keeps its table of contents, and AI-continued branch content can be narrated the same way.
Can I listen without a network connection?
System TTS works offline once an offline voice pack is installed. For online voices, download the chapter on Wi-Fi first — cached chunks play without a network, and uncached chunks queue for synthesis when you are back online.
Do listening and reading positions drift apart?
Not while you listen with the reader open: the page follows the narration (scroll and page modes both), and that position is written into your reading progress. The “Text” button on the listening screen returns you to the reader, and narration starts from the chapter your reading position is in.
Which online voice services are supported?
Any OpenAI-compatible speech endpoint (custom API host + key + voice name) — including providers like Alibaba Model Studio, MiniMax, and xAI. Keys are encrypted on device and requests go directly to the provider.
Now on Google Play
Put the next book in your ears
The Android app is live on Google Play; for the iOS beta we send one short note when the next wave opens. You can also email [email protected].