How do I make an AI audiobook in ChatGPT?

Yes, with Creative Claw and a digital voice. Paste a chapter, pick a voice and a reading style, and it returns the audio with word timings. Our 2:19 chapter sample cost 30 credits.

Our sample ran 2 minutes 19 seconds for 30 credits with Gemini 3.8 Flash TTS, about 13 credits a minute, or roughly 800 credits ($8) per finished hour at that pace. Human narrators charge about $100 to $500 per finished hour.

Not through ACX, which requires human narration unless authorised. Amazon’s Virtual Voice (an invite-only KDP beta) is the AI route there. Use Creative Claw narration where digital voices are accepted, for samples, or to audition a script before hiring a narrator.

Gemini 3.8 Flash TTS has 38 curated voices, such as a soft storyteller and a low dramatic narrator, plus Google’s wider library; ElevenLabs voices are also available. You can direct the style, such as hushed and suspenseful.

Yes, with your consent and a recording of your own voice: Creative Claw can clone it and read your text. A cloned voice is still digital narration under store rules.

Gemini 3.8 Flash TTS takes up to 10,000 characters per request, about 15 minutes at our pace, so narrate a book chapter by chapter.

Usually. Our town name was read correctly in English; check names in the transcript, and respell a difficult name the way it sounds if it comes out wrong.

Yes. The same voice read our opening in Spanish for 8 credits; Gemini’s speech model covers many languages. Have a native speaker check the translation and pronunciation.

Yes. ElevenLabs v4 voices up to 10 characters in one request, each line with its own emotion. Our three-voice scene (a narrator, Ada and an old harbourmaster) ran 51 seconds for 16 credits.

Yes. Write them into the script in brackets, such as [waves crashing], [wind howling], [paper tearing] or [thunder], and ElevenLabs v4 performs them in the same audio as the voices. For exact timing or layered ambience, generate the effect separately and mix it.

Put a tag before the words: [whispers], [shouting over the wind], [trembling], [gasps], [horrified whisper]. Our transcript check even caught the gasp. Audition important lines, since tags are guidance, not exact control.

In our tests: Eve for a whispered thriller, Amelia for tender romance, Rex Thunder for epic fantasy, Sia for a sarcastic cozy mystery, Austin for Southern gothic horror, Paulette for quiet literary grief, Spuds Oxley for bedtime stories and Brian for Latin American Spanish. Each 16- to 24-second excerpt cost 5 to 7 credits; listen to all eight above.

ElevenLabs v4 for emotion tags, several characters and effects in the script; Gemini 3.8 Flash TTS for long, even narration at about 13 credits a minute. Many authors use v4 for samples and trailers and an even voice for the full book.

Yes. Ours plays over the trailer shots with each speaker’s lines captioned in their own colour, cut with HyperFrames. Creative Claw’s video agent builds edits like this in the cloud from a description (about 25 to 450 credits, depending on length; see creativeclaw.co/video-agent).

Transcribe the audio (about 2 credits for our sample) and compare it with your text. Our transcript matched every word of chapter one.

Yes. KDP’s content guidelines ask you to disclose AI-generated content, including text, images (cover and interior art) and translations, when you publish or republish. Content an AI tool created counts as AI-generated even if you edited it a lot. AI-assisted work (you made it, AI helped edit or brainstorm) does not need disclosure.

Not through ACX. ACX’s submission requirements (April 2026) say audiobooks must be narrated by a human unless otherwise authorised, and unauthorised text-to-speech is prohibited. Amazon’s own Virtual Voice, an invite-only KDP beta, is the authorised AI route.

Spotify for Authors accepts digital voice narration from providers “like Google Play Books, ElevenLabs, and more” and adds a disclosure line. Google Play Books accepts uploads with a synthesized-voice narrator. Kobo Writing Life accepts AI narration credited to a synthesised voice (as reported). Apple Books offers only its own digital narration. Check each store before you narrate a whole book.

Your text stays yours. In the US, the Copyright Office says prompts alone do not make an image copyrightable; your own selection, arrangement and edits can be. Keep your drafts. This is not legal advice.

No. The FTC rule in force since October 21, 2024 bans fake reviews and testimonials, including AI-generated ones, with civil penalties for knowing violations. Quote your own book, or real reviews you have permission to use.

On YouTube and TikTok, label realistic AI-generated video with the platform’s AI disclosure. A stylised or clearly illustrated trailer usually is not covered; check each platform’s current rules when you post.

Open chatgpt.com/plugins/creativeclaw and select Install plugin, or search for Creative Claw in ChatGPT’s plugin directory, then sign in. New accounts get 100 free credits, no card.

No. Creative Claw is pay as you go: $10 buys 1,000 credits that never expire, and new accounts get 100 free credits. You pay per book, not per month.

Yes. Paste a chapter or your blurb into ChatGPT for narration and trailer scripts, and attach your cover so trailers, quote cards and translated editions match it.

No. It makes the files (covers, trailers, narration, quote cards); you upload them to KDP, your audiobook distributor or your social accounts.