ElevenLabs Text to Speech Persian: Farsi Voices Guide

Yes. ElevenLabs offers Persian text to speech on its official Persian page, with a demo in Persian script, the shared voices Jessica, Laura, Alice, Bill, and Brian, and expressive bracketed cues (checked 5 Nov 2026). Because Persian is outside Multilingual v2 and Flash v2.5, use Eleven v3 or Eleven v4.

Reading the Persian page in order

The Persian page opens with a free Persian text-to-speech offer and a demo box. On 5 Nov 2026 that demo sample was written in Persian script, with a character counter beside it, and it carried the same bracketed direction cues seen on the other language pages: [sarcastically], [giggles], and [whispers]. The opening of the page also speaks of the rich storytelling tradition of Iranian culture. That framing is a useful guide to tone: Persian audio works best when the script is written to be heard as a story, with clear scenes and pauses, rather than as a dense block of information.

Below the demo, the voice list names Jessica, Laura, Alice, Bill, and Brian. The page does not describe those five as Persian-native voices, so this rewrite does not present them that way. Judge them on Persian lines, one voice at a time, and listen for sentence flow, emphasis, and the handling of longer Persian sentences. A voice that sounds bright on a short greeting can still lose its shape over a long passage, so the demo is a starting point, not a final verdict on a voice.

Why the older model lists do not carry Persian

Persian is not in the 29-language Multilingual v2 list from the 5 Nov 2026 model check, and it is not added by Flash v2.5, which covers those 29 plus Hungarian, Norwegian, and Vietnamese. That is the whole reason the model step matters for Persian: a plan that assumes the older lists cover every language on the site will fail here. Choose Eleven v3 or Eleven v4 for Persian, and check the model name in the dashboard before generating a long passage, because a model change after generation means paying for the same text twice.

The expressive side follows the same model rule. Bracketed cues such as [sarcastically] belong to the v3 and v4 models, a point set out in our emotion tags guide. If a cue is read aloud as plain text, or ignored, the model is the first place to look. Our Multilingual V2 guide also explains the older model in full, including its 29-language scope and its 10,000-character request ceiling, which helps explain why a Persian job should not be built on it.

A Persian test script that respects the demo

Write the test in Persian script, exactly as the audience will read it, and keep it short. A test of two or three sentences is enough to hear whether a voice holds Persian rhythm, and it costs only a small number of credits. Read the test script aloud yourself first, and mark the places where a speaker would naturally pause: after an introduction, before a new idea, and at the end of a quoted line. Those marks become line breaks in the script, and line breaks are the simplest way to give a generated voice room to breathe.

1. Draft a short Persian passage in Persian script, with one idea per sentence.
2. Select Eleven v3 or Eleven v4, then paste only that passage into the demo.
3. Listen to Jessica and Alice first, then to Bill, on the same passage, and note which voice keeps the story tone.
4. Add a single cue, such as [giggles], and listen before adding any second cue.
5. Once a voice wins the short test, generate the full script in sections and check each section before moving on.

One practical warning about pasting: Persian script is written right to left, so a script copied from a left-to-right document can arrive with punctuation in the wrong place. Check the first and last characters of each pasted paragraph, and check that question marks and full stops sit where the sentence ends. A misplaced full stop is a small typing fault that sounds like a large voice fault, because the voice will pause in the wrong place.

Six uses, one storytelling thread

The Persian page lists Persian AI Agents, Podcasts, Audiobooks, Gaming, Video voiceovers, and Accessibility. Those six uses share one thread: each one asks a listener to follow a spoken story or instruction without seeing the text. An agent must answer in a helpful tone, a podcast must hold attention across an episode, an audiobook must keep a single voice steady across chapters, a game line must land in a few seconds, a video voiceover must match the picture, and accessibility audio must be clear above all else. The script for each use should be written for that listening situation, not copied from a written article.

Persian story audio: podcasts and audiobooks

For podcasts and audiobooks, keep one voice for a whole series and give each chapter a short spoken heading. A listener who pauses mid-chapter needs that heading to find the place again, and a voice change mid-series is more distracting than a small wording change. Write chapter breaks as pauses, not as headings alone, because a heading on paper does not create a pause in audio unless the script creates one.

Persian spoken help: agents and accessibility

For agents and accessibility audio, put the key fact first and keep sentences short. A listener who asks for help cannot scan back over a long spoken answer, so the first sentence should give the answer, and later sentences should give the detail. Accessibility narration should favour a steady, clear voice over a dramatic one, even when the same script could support a more theatrical reading for a podcast.

Persian picture audio: gaming and video voiceovers

For gaming and video voiceovers, write to the length of the scene. A game line that runs past the animation looks unfinished, and a video narration that runs past a scene change confuses the viewer. Mark emotional lines with a single bracketed cue and leave the rest of the script plain, so the cue stands out when it is used. Too many cues in a short passage make every line sound exaggerated, and the passage loses its shape.

What Persian audio costs in credits

Persian has no separate price on the official pricing page, and all prices are in USD. Free gives 10,000 credits a month at $0, Starter is $6 per month after a $1 first month shown until 8 Nov, Creator is $22 per month after an $11 first month, and Pro is $99 per month (all checked 5 Nov 2026). Our ElevenLabs pricing guide gives the full plan table, and our ElevenLabs languages guide explains how the 90-plus language total on v4 sits beside the smaller lists on older models.

At about one credit per character, a 2,000-character announcement costs roughly 2,000 credits, and a 20,000-character lesson series costs roughly 20,000 credits. Count the characters in the script before generating, including spaces and bracketed cues, because a heavily cued script costs more than the same words without cues. For a regular Persian series, Starter is the first paid step, and Creator is the step for a weekly show or a growing audiobook catalogue.

When Persian audio will disappoint

• A long, unbroken Persian paragraph with no pauses will sound flat, whatever voice is chosen.
• A script pasted with punctuation in the wrong place will pause in the wrong place, and the voice will take the blame.
• A voice chosen on a greeting alone may not hold up over a full chapter, so test a chapter section first.
• A job built on Multilingual v2 or Flash v2.5 will not serve Persian, because Persian is outside both lists.
• An unproofed script wastes credits, because every regeneration of a corrected passage costs credits again.

Who should NOT use ElevenLabs for Persian

Skip ElevenLabs for Persian if the job needs a voice that the official Persian page itself presents as Persian-native, because that roster is not shown there. Skip it also for a recording that must be certified as a native-speaker performance for legal or broadcast purposes, since a generated voice is not a certified human recording. A third reason to skip it is a hard requirement to use Multilingual v2 or Flash v2.5: Persian is outside both lists, so that requirement rules ElevenLabs out for Persian unless the model choice can change.

Frequently asked questions

Does ElevenLabs do Persian text to speech?

Yes. The official Persian page presents Persian text to speech with a demo in Persian script, checked 5 Nov 2026. The page speaks of the rich storytelling tradition of Iranian culture, and it lists agents, podcasts, audiobooks, gaming, video voiceovers, and accessibility as uses.

Which model should I use for Persian?

Use Eleven v3 or Eleven v4. Persian is not in the 29-language Multilingual v2 list or the 32-language Flash v2.5 list from the 5 Nov 2026 model check, so neither older model is the right pick for Persian. The v3 and v4 models also support the bracketed cues shown in the Persian demo.

Are the Persian voices native Persian voices?

Not on the evidence of the page. The Persian page on 5 Nov 2026 lists Jessica, Laura, Alice, Bill, and Brian, the shared five voices also shown on other language pages, without a Persian-native label. Test those voices on Persian lines and judge by ear, rather than assuming any one of them is a Persian-native voice.

How do I paste Persian text without errors?

Paste the script in Persian script, then check the first and last characters of each paragraph and the position of full stops and question marks. Persian script runs right to left, so punctuation copied from a left-to-right document can land in the wrong place. A short test generation will also reveal a misplaced pause before a long script is generated.

Is Persian text to speech free on ElevenLabs?

Yes, up to the Free plan limit. The Free plan costs $0 and includes 10,000 credits a month (checked 5 Nov 2026), which covers short Persian tests and announcements. Longer or regular Persian work needs a paid plan, starting with Starter at $6 per month after a $1 first month.

Can I use Persian audio for commercial projects?

Yes, on paid plans. Starter and higher plans include a commercial licence, so Persian announcements, lessons, and video narration made on those plans can be published commercially. The Free plan is for testing, so move to Starter before publishing paid client work or monetised content.

Does the Persian page cover Dari or Tajik?

No separate Dari or Tajik section is shown on the Persian page checked on 5 Nov 2026. The page is presented as Persian, and this rewrite does not claim separate coverage for Dari or Tajik, because that coverage is not shown on the page. Test any Dari or Tajik lines separately and judge them by ear before publishing.

Sources checked

• ElevenLabs Persian text to speech page — demo, voice list, and use cases, checked 5 Nov 2026.
• ElevenLabs pricing page — Free, Starter, Creator, and Pro prices and credits, checked 5 Nov 2026.
• ElevenLabs models documentation — model language counts used for the v2 and Flash v2.5 check, checked 5 Nov 2026.