LyricWinter vs ElevenLabs: Keep v4, Give Each Character a Voice

If you've listened to a fic in a read-aloud app, you've probably heard one voice read every character. LyricWinter gives each character their own voice. LyricWinter Studio is where you make the audio. It runs in your web browser and keeps each story as a project, with one section per chapter. It's for readers who want an MP3 of a fic with lots of dialogue for the drive or the gym, and for writers who want to hear a chapter before they post it.

We make LyricWinter. The ElevenLabs facts here come from its own website and help pages, read on October 1, 2026. Both samples below were made in LyricWinter Studio.

Listen: one scene on ElevenLabs v4, then on two engines

In "Floor Nineteen," a maid of honor is stuck in a hotel elevator with her ex, who left for Lisbon. Bev from the front desk has opinions over the intercom. Studio found the four speakers and picked a voice for each. In this first version, every voice runs on ElevenLabs v4.

Auto emotion works by writing acting notes. For v4, these notes are audio tags, such as [whispers]. Studio wrote them and placed the sound effects for you. The sound effects are free.

Next is the same scene, with the same words and the same four voices. Tamsin and Dario stay on ElevenLabs v4. The Narrator and Bev move to Inworld TTS 2, a different voice engine. TTS is short for text to speech.

Mixing engines lets you control cost. You pay for audio in LyricWinter words, credits you spend when you make audio. For English stories, each engine uses about one LyricWinter word per word of story. That's 1x. From October 12, 2026, ElevenLabs v4 becomes 3x, three times as much. The change happens at midnight UTC, a world standard time. In the US, that's still October 11.

Tamsin and Dario speak about 40 percent of the scene's words, so after the change, only their lines cost 3x. That makes the second version cost about 40 percent less than the first, not counting the audio tags. The cost table shows the same math for a whole chapter.

What you get in LyricWinter

ElevenLabs v4 is one voice engine. LyricWinter is everything from the link you paste to the file you download: the import, the voices, the engines, the sharing, and the audio you own.

Start from a fic link, without the author's notes

There are two ways to bring a story in. When you start a new project, the Import menu takes an AO3 work, an AO3 chapter, or an EPUB file (an ebook file). An AO3 work becomes a project with one section per chapter. Studio shows you the chapter list first, and nothing is created until you confirm.

You can also import into a section you already have. Open its three-dot menu and choose Import story, then From link. Paste a link from AO3, FanFiction.net, SpaceBattles, or Substack, or paste the text yourself. An AO3 link lets you pick Current chapter or Full work. The import leaves out the summary, the author's notes, and the info pages at the start and end, so the audio starts with the story.

Our AO3 to MP3 guide walks through it, and AO3 text to speech with a voice per character shows how to hear any AO3 chapter as a cast.

ElevenLabs Audiobooks, by comparison, takes an EPUB or PDF upload or pasted book text, according to its docs.^[1] If you download a fic from AO3, the file adds a page with the rating, tags, and summary, and it keeps the author's notes.^[2] Read those notes on AO3 before you listen, and leave kudos if you enjoyed the fic. If the fic is someone else's, keep your audio private unless the author says yes.

An engine for each character

LyricWinter Studio connects 12 voice engines from 7 providers: ElevenLabs v4, Inworld, Fish Audio, Cartesia, OpenAI, Smallest AI, and F5-TTS. Each character has its own engine menu, so one chapter can use several engines at once.

Auto emotion tells each voice how to say its lines, in a form each engine understands: audio tags for v4, instructions for Inworld TTS 2, and emotion settings for Cartesia Sonic 3.5. To hear one voice on several engines before you choose, use Compare voice engines. It plays the same short sentence on each engine and costs words like any other audio.

Clone a voice once, use it on up to 9 engines

A voice clone is an AI copy of a real voice, such as your own. Upload a clear clip of one person talking, 10 to 30 seconds long, or record one in your browser. Then confirm you have permission to use the voice. Once its sample is ready, the voice works on up to 9 engines, ElevenLabs v4 included. The first time you use it on an engine, LyricWinter sets it up for you, so you only upload once.

A free cast, with new voices designed for it

Finding the characters is free. LyricWinter finds everyone who speaks, adds a Narrator for the text outside the quotes, and picks a voice for each. If no voice fits a character, LyricWinter designs a new one from their lines. You can also design a voice by describing it in writing. Designing voices is free too.

Pin a voice to reuse it

To reuse Dario's voice, open his Voice settings and audio and choose Pin as default. After that, any new project with a character named Dario gets the same voice and engine. Inside one project, every character keeps their voice from chapter to chapter. See our voice pinning guide.

You own the audio you make

You own the audio you make, so you can use it however you like, as long as you have the right to use the voices and the text. That includes posting it, putting it in a podcast or a video, or selling it, on every plan. Export final mix saves each section as an MP3 or WAV audio file, or as SRT captions (a subtitle file). Projects and sections stay private until you create a listening link.

Listening links and a player for your website

Create listening link gives you one link to send to a friend, a beta reader (someone who checks your story before you post it), or your group chat. The link stays current as you edit, and Stop sharing turns it off. Listeners see the full text with speaker names, the passage being read is highlighted, and they can start from any passage. If you have a website, choose Prepare audio embed, then Copy embed code. Paste that code into your page to add a player. The player links back to the page where listeners read along.

Words that never expire

You pay by the word for the audio you create. Paid words never expire, and Studio charges only for the lines it finishes. Word packs are one-time purchases, so you don't need a subscription. Plans are optional. The cost table below shows a chapter and a novel.

What is the difference between LyricWinter and ElevenLabs?

LyricWinter vs ElevenLabs at a glance, as of October 1, 2026, from each company's own pages and docs.

Sources for the ElevenLabs column: its Studio docs,^[3] its Audiobooks docs,^[1] its Character Casting announcement,^[4] and its terms of service, last updated March 31, 2026.^[5]

How do you mix ElevenLabs v4 with other engines in LyricWinter?

In LyricWinter Studio, each character has a card (a small box with their name and voice) in the Characters panel. You pick their engine there. You don't need an ElevenLabs account: ElevenLabs v4 is one of the choices, and you pay with LyricWinter words.

• Studio starts new characters on v4. As of October 1, 2026, Studio puts a new character on ElevenLabs v4 whenever their voice supports it, so you hear v4 first. A pinned voice keeps the engine you pinned with it. From October 12, v4 costs 3x words, and its label in Studio adds "3× words". To pay 1x, switch the character to another engine before you create audio (steps 3 and 4 below).
• Which voices can use v4. Any LyricWinter voice with a finished voice sample: one you uploaded, one designed for a character, or one from LyricWinter's shared library of ready-made voices and voices other users created. The first time a voice runs on v4, LyricWinter sets it up for you.

Steps in LyricWinter Studio

1. Open LyricWinter Studio in your browser and sign in with Google or Apple. The link opens New project when you're signed in. If you don't see it, tap New project. To start blank, type a project name, add a New section, and paste your chapter. To bring in a story, open the Import menu and choose AO3 work, AO3 chapter, or EPUB file.
2. Choose Manual voice assignment. On a phone or other narrow screen, it's under the small arrow on the Create audio button. Studio finds the characters and suggests a voice for each. You can change any of them before you make audio. This step is free and makes no audio.
3. In the Characters panel, tap the engine name on a character's card, such as ElevenLabs v4. Pick another engine.
4. A box asks Change voice engine? Confirm, for example with Change to Inworld TTS 2. The change applies to that character in the whole project.
5. The Create audio button shows how many words the audio will use. Check that number, then click Create audio.

To keep the engines Studio picked, skip steps 3 and 4. To make v4 whisper, shout, or cry, see our ElevenLabs v4 audio tags guide, which covers the tags and Auto emotion.

How much does a mixed ElevenLabs v4 cast cost on LyricWinter?

A 10,000-word chapter makes about an hour of audio. On a 1x engine, it costs about $4. ElevenLabs v4 costs the same until October 12, 2026 (midnight UTC), then three times as much. The table uses $1 per 2,500 words and v4's later 3x price. It counts the story text only, without audio tags.

Is LyricWinter right for you?

Yes, if you want a fic, a web serial (a story posted online in parts), or your own chapter performed by a cast instead of one voice. You can start from a link or pasted text, and you can start free. The free plan includes 2,500 words a month, about 17 minutes of audio on a 1x engine. That's enough to hear a scene before you buy words. The first sample above is all ElevenLabs v4. Once v4 costs 3x, use it where it matters most and a 1x engine elsewhere.

Two things to know before your first story:

• Link import covers four story sites. AO3, FanFiction.net, SpaceBattles, and Substack, plus direct links to documents such as a PDF. For other sites, or an AO3 work only logged-in members can see, paste the text.
• Studio is in beta and exports one file per section. Beta means it's an early version. Import a work as a project, and each chapter becomes its own section and its own file. On the free plan, Studio makes audio for up to 30,000 characters per section. Here, characters means letters, spaces, and punctuation. That's about 5,000 words. The limit is 90,000 characters on the MyWinter plan and 150,000 on the UltraWinter plan. Paid plans split a longer chapter into sections for you.

How LyricWinter works explains the steps from text to audio. Listening to fics? Start with LyricWinter's fanfiction reader and our guide to listening to fanfiction with a full cast. Comparing tools? See the best text to speech for fanfiction and the best AI audiobook generator for multi-voice audio. Writers checking their own chapters can read hear your chapter before you post.

LyricWinter vs ElevenLabs: FAQ