Quick summary: ElevenLabs gives every new account a Free plan with 10,000 credits a month, which is enough for roughly 10 minutes of AI voiceover in Text to Speech. This guide walks through signing up, picking a voice and model (including the new Eleven v4), tuning the sliders, downloading your audio, and the free-plan rules you need to know before you publish anything.
Table of contents
- Create a free account
- Open Text to Speech and paste your script
- Pick a voice that works on the Free plan
- Choose a model: Eleven v4, Multilingual v2 or Flash
- Set Stability and Similarity
- Direct the delivery with audio tags
- Generate, regenerate for free, and download
- Make your own voice with Voice Design
- Try long-form narration in Studio
- Free vs paid plans
- FAQ
Step 1. Create a free account
Go to the ElevenLabs sign-up page and register with email, Google or Apple. You don't pick a plan: ElevenLabs puts every new account on the Free plan automatically, and you can check your remaining credits any time on the subscription page in your profile menu.
One practical tip: sign up on your normal home or mobile connection, not a VPN. ElevenLabs' help center says its abuse system can block free-tier use with an "Unusual activity detected" message when it sees a shared or flagged IP address.
Step 2. Open Text to Speech and paste your script
Choose Text to Speech in the left sidebar. Type or paste your text into the large input box. The official quickstart says you can enter up to 5,000 characters per generation there, so split a long script into chunks. On the website, each character usually costs one credit, which is why 10,000 free credits works out to about 10,000 characters a month.
Step 3. Pick a voice that works on the Free plan
Open the voice selector in the Settings panel on the right. You can use the default voices or search the Voice Library, which the quickstart describes as having more than 3,000 voices you can filter by language, accent, age and use case. Two things to watch on a free account:
- Some library voices show "This voice is not available for free users" because their owner restricted them.
- Voices with a credit multiplier cost more. A 2x voice uses twice the credits for the same text.
For non-English scripts, ElevenLabs recommends a native voice for that language for the most natural result.
Step 4. Choose a model: Eleven v4, Multilingual v2 or Flash
Under the voice, open the model menu. The models you'll see most often:
- Eleven v4: the newest and most emotive model, with 90+ languages and natural multi-speaker dialogue. ElevenLabs calls it a net upgrade over v3 for almost every case.
- Eleven Multilingual v2: the docs call it the most stable choice for long-form narration.
- Eleven Flash v2.5: built for very low latency, mainly for apps and voice agents.
Voice choice matters most, then the model, then the sliders, so test two or three voice and model pairs with a short sentence before you spend credits on the full script.
Step 5. Set Stability and Similarity
The docs suggest starting around 50 for Stability and 75 for Similarity, with Style Exaggeration left at 0. Lower Stability gives a more expressive, varied read. Higher Stability sounds steadier but can turn flat. If you're on Eleven v4, you'll only see Stability and Similarity, because v4 drops the Style and Speed sliders and doesn't support SSML.
Step 6. Direct the delivery with audio tags
With Eleven v4 and v3 you can write short tags in square brackets right inside your text, such as [whispering], [laughing] or [sighs]. Put the tag just before the line it should affect. ElevenLabs notes that tags still aren't followed perfectly, so expect to try a line a couple of times.
Step 7. Generate, regenerate for free, and download
Click Generate speech. Credits are charged when you generate, not when you download. If you don't like the take, you can get two free regenerations as long as the text, voice and model stay the same, it's within two hours of the first take, and you haven't refreshed the page. You can still move the sliders.
To save the file, use the download button after generating, or open the History tab and choose MP3 (128 kbps) or WAV.
Step 8. Make your own voice with Voice Design
Can't find the right voice? Go to Voices, then My Voices, then Add a new voice, and pick Voice Design. Describe the voice in plain words, for example "A calm female narrator, 30s, warm and clear." You get three options and pay only the credits for the preview text once. Saving one uses a voice slot. Instant voice cloning of your own voice needs a paid plan.
Step 9. Try long-form narration in Studio
Studio is ElevenLabs' editor for audiobooks, narrated videos and longer scripts. You can upload EPUB, PDF, DOCX, TXT or HTML files, or import a web page by URL, assign voices to sections, regenerate single paragraphs, and export MP3 or WAV. The Free plan includes 3 Studio projects, and free video exports carry a watermark.
Free vs paid plans
| Feature | Free | Starter | Creator | Pro |
|---|---|---|---|---|
| Monthly price (USD) | $0 | $6 | $22 ($11 first month) | $99 |
| Credits per month | 10,000 | 30,000 | 121,000 | 600,000 |
| Commercial license | No | Yes | Yes | Yes |
| Instant Voice Cloning | No | Yes | Yes | Yes |
| Professional Voice Cloning | No | No | Yes | Yes |
| Unused credits roll over | No | Yes (up to 2 months) | Yes (up to 2 months) | Yes (up to 2 months) |
| Studio projects | 3 | 20 | At least 20 | At least 20 |
Prices are from the official ElevenLabs pricing page as of October 2026 and exclude taxes. Annual billing costs 10 months' worth.
FAQ
Is ElevenLabs really free?
Yes. The Free plan costs $0 and renews 10,000 credits each month. Unused free credits don't carry over to the next month.
Can I use free ElevenLabs audio on YouTube?
Only for non-commercial use. ElevenLabs says the Free plan has no commercial license, and if you publish free-plan content you must credit it by putting "elevenlabs.io" or "11.ai" in the title. For a monetized channel, ads or client work, you need a paid plan, and only audio made while subscribed counts.
Why do I see "Free Tier usage disabled"?
The abuse filter flagged your connection, often because of a VPN or a shared IP at school, work or on mobile data. ElevenLabs says the fix is either a paid plan or a unique IP that hasn't been flagged.
Related guides on this blog
- How to Use Nano Banana in Gemini for Free (2026 Guide)
- How to Use NotebookLM (Gemini Notebook) for Free in 2026
- How to Use Microsoft Copilot for Free in 2026 (Step-by-Step)
Which voice or model gave you the best result? Tell me in the comments.
Screenshots at the end of this post: ElevenLabs (official documentation images).
Sources
- ElevenLabs Pricing
- ElevenCreative quickstart (ElevenLabs Docs)
- Text to Speech guide (ElevenLabs Docs)
- Eleven v4 (ElevenLabs Docs)
- Voice Design (ElevenLabs Docs)
- ElevenCreative Studio (ElevenLabs Docs)
- Can I publish the content I generate on the platform? (ElevenLabs Help)
- Free regenerations in Text to Speech (ElevenLabs Help)
- "Free Tier usage disabled" message (ElevenLabs Help)
댓글
댓글 쓰기