Persian Text to Speech
Natural Persian voice from any Farsi text — 50+ AI voices, free MP3.
50+ Neural Voices — Farsi, Dari & RTL Script
From Rumi's ghazals to Tehran evening news, Farsi carries a poetic cadence that flat robotic voices destroy — these neural voices preserve it. The library includes 53 neural speakers trained on modern Iranian pronunciation — correct ezafe linking, natural word-final stress, and smooth handling of the Perso-Arabic script with all four uniquely Persian letters (پ چ ژ گ). Choose Farid for an authoritative male reading or Dilara for a warm, conversational female delivery, then download your file free.
The catalogue covers tones suited to audiobook narration, podcast voiceover, Farsi language learning, business presentations for the Iranian market, and heritage-speaker practice. Think of it as a Farsi text to speech reader that handles everything from short captions to full manuscripts. Right-to-left rendering is handled automatically — just paste your text and the engine reads it with authentic Tehran intonation. First 1,000 characters free, no account required.
- 53 voices — all Neural tier
- Farsi, Dari-adjacent & RTL-ready
- Adjustable speed & pitch
- Download MP3, WAV, FLAC, OGG
- Free — 1,000 chars, no signup
Persian AI Voice — Voice Samples
Click to preview · 53 native voices total
These are 4 featured speakers out of 53 Farsi TTS voices. Browse the full list on the voices page — filter by fa-IR.
Persian Pronunciation & Phonetic Highlights
Eight phrases that show how the engine handles Farsi phonetics — from everyday greetings to classical poetry. Click play and follow the transliteration.
What Makes Farsi Phonetics Distinctive
- Word-final stress — unlike English, most words in Farsi carry stress on the last syllable. The engine reproduces this pattern correctly, so ketāb (book) lands on the second beat, not the first.
- Ezafe linking — a short -e or -ye connects a noun to its modifier (ketāb-e man = my book). This vowel is rarely written in standard orthography, yet the reader inserts it where grammar requires.
- Velar and uvular fricatives — the sounds kh (خ) and gh (غ) are produced deep in the throat. Both are modelled accurately, which matters for natural output and for learners who need a reliable Persian reader as a pronunciation reference.
Farsi Script — Formatting & Conventions
Small formatting choices change how the engine reads your input. Four conventions worth knowing for Farsi:
Numbers
۱۲۳ → "sad-o-bist-o-se" — paste Eastern Arabic numerals or Western digits; the engine converts both correctly. Currency amounts like ۵۰٬۰۰۰ تومان are read as spoken Farsi, including the hezār (thousand) convention.
Diacritics
Standard Farsi text omits short vowels. The engine infers zir, zebar, pish from context — no diacritical marks needed. Add a tashdid (ّ) only when disambiguation matters, for example in شُکر (thanks) vs شِکر (sugar).
Dates
۱۴۰۴/۱/۲۳ — Solar Hijri calendar format (year/month/day). The engine reads it naturally: "bist-o-sevvom-e Farvardin-e hezār-o-chahār-sad-o-chahār". You can also paste Gregorian dates; both parse correctly.
Mixed Scripts
Persian text often embeds Latin brand names or English terms. The engine switches direction on the fly — iPhone 15 inside a Farsi paragraph reads in the correct hybrid order with no manual tagging required.
Use Cases: Persian Voice in Action
Content Creation & Voiceover
Add a Farsi narrator to YouTube videos, podcasts, and social reels. Pick a voice that fits — authoritative for tech reviews, warm for lifestyle content — and export the audio file for Premiere, DaVinci, or CapCut. The Iranian diaspora audience on YouTube alone is massive; a natural-sounding narrator makes your channel stand out.
Farsi Language Learning
Slow the playback to 0.75x and follow along with common phrases, verb conjugations, or vocabulary lists. Hearing correct stress placement and vowel quality from a native speaker accelerates retention, especially for the ezafe pattern that textbooks rarely illustrate with audio.
Audiobooks & Persian Literature
Turn manuscripts into audiobooks — from the ghazals of Hafez and masnavi of Rumi to contemporary Iranian novels. The classical register sounds best at a slower tempo (0.85x); modern prose reads naturally at default speed. Use Dialog Mode to assign different speakers to characters.
Pronunciation Practice & Heritage
Heritage speakers who grew up outside Iran can use the Farsi reader to tune their accent by repeating key phrases at reduced speed. The engine handles velar fricatives (kh, gh), the glottal stop (hamze), and colloquial Tehrani contractions, giving you a reliable reference when a live tutor is not available.
Persian Text to Speech — How It Works
Three steps to generate Farsi audio online. No software, no signup.
Paste or type your Farsi text
Type directly or paste up to 1,000,000 characters. Upload DOCX, PDF, or SRT files. Works with any right-to-left input — articles, scripts, poetry, study notes.
Choose a voice
Pick from 53 neural speakers. Filter by gender, then adjust speed and pitch to fine-tune the reading. Try Farid for formal narration or Dilara for casual delivery.
Listen & download free
Click Convert to Speech, preview the result, and download as audio. First 1,000 characters free — no account needed, no watermark on any plan.
Persian Language Spotlight
What sets Farsi apart — and how the engine handles it.
Farsi, Dari & Tajik
All three are varieties of one language. Iranian Farsi (70–80 million speakers) is the standard modelled here. Dari (Afghan variant, 12 million) shares the written form but differs in vowel quality and vocabulary. Tajik (8 million) uses Cyrillic script. If your text is written in standard Perso-Arabic orthography, the engine will read it naturally regardless of regional origin.
Nastaliq Calligraphy
The calligraphic style نستعلیق is the visual identity of written Farsi — distinct from the Naskh style used in Arabic. While the engine processes plain Unicode text (not images), knowing that your audience expects Nastaliq rendering matters for any visual pairing you create around the audio.
Ezafe & Unwritten Vowels
The ezafe construction (-e / -ye) links nouns to adjectives and possessors but is almost never written. The engine infers ezafe placement from syntax: ketāb-e bozorg (the big book) reads correctly from کتاب بزرگ alone, with no extra diacritics needed.
Perso-Arabic Script + 4 Extra Letters
Farsi uses the Arabic alphabet plus four letters that Arabic does not have: پ (p), چ (ch), ژ (zh), گ (g). The engine maps all four to their correct phonemes — critical for words like gol (flower) and cherā (why).
Frequently Asked Questions
The voices are trained on Iranian Farsi pronunciation. Because Dari shares the same written script, you can paste Dari text and the engine will read it correctly in most cases. Vowel quality and some loanwords will lean Iranian rather than Kabuli, but grammar, consonants, and script rendering are identical. For content specifically targeting Afghan audiences, listen to a preview and adjust phrasing where the accent matters.
Yes. The SpeechGen API accepts any fa-IR voice identifier and returns audio in the format you specify. Send a POST request with your text, choose a voice like fa-IR-FaridNeural, and receive a download link. Integrate Persian TTS into your own app or workflow — a free tier is available for testing.
They are two names for the same language. "Persian" is the English exonym (like "German" for Deutsch); "Farsi" is the endonym used by native speakers (فارسی). In academic and diplomatic English, "Persian" is more common; in everyday tech usage, "Farsi" dominates. This page covers both — the voices, the script, and the phonetics are identical regardless of which label you prefer.
Yes. The first 1,000 characters are free with no account, no watermark. Create a free account and you receive an additional 3,000 characters per day for seven days. Paid plans raise monthly limits and unlock extras like bulk export, longer scripts, and commercial licensing on every tier.
Yes. Standard Farsi text almost never includes diacritical marks for short vowels. The engine uses context and a built-in lexicon to resolve ambiguities — for example, distinguishing shokr (gratitude) from shekar (sugar) based on surrounding words. Add a tashdid or kasra only when the default reading sounds wrong, which is rare in normal prose.