Text to Speech Languages: One Voice, 30+ Languages Fish Audio

One engine reads them all: voices on this site are not locked to a single language, and a voice that carries your English videos also reads the Arabic, German or Thai version of the same script. That changes what a multilingual channel or product needs — not a voice per language, but a script per language written the way each language wants to be read.

This page collects every language guide on the site. Each one covers the same three things for its language: what is different in the script, what to check before generating, and what a project costs.

Guides by language

• Arabic text to speech — standard Arabic and dialect voices for MENA audiences, with the register and script notes that matter for ads and dubbing.
• Persian text to speech — RTL script, numbers, names and formal or conversational Farsi for Persian-speaking audiences.
• Spanish text to speech — neutral, Castilian and Latin American Spanish voices, and how to pick the one your market expects.
• Russian text to speech — Cyrillic reads directly; the work is stress marks, register and sentence length.
• Turkish text to speech — spelling, suffixes and loanwords, with the pacing that keeps agglutinative sentences breatheable.
• Indonesian text to speech — formal and casual Bahasa Indonesia, and when each register is the right one.
• Thai text to speech — no spaces between words, tones that decide meaning, and particles that set the register.
• German text to speech — compounds, umlauts and the sharp s, and the du or Sie decision that runs through the whole script.
• Polish text to speech — diacritics that decide words, digraphs, penultimate stress and the Pan or Pani register.
• French text to speech — accents, liaison, silent endings, register and natural French pacing.
• Portuguese text to speech — Brazilian and European Portuguese vocabulary, rhythm and market choices.
• Japanese text to speech — kana and kanji readings, honorifics, names and Japanese pacing.
• Korean text to speech — Hangul, sound changes, speech levels, loanwords and numbers.
• Hindi text to speech — Devanagari, Hinglish, numbers, names and formal or conversational register.

What changes from language to language

• The register system. German splits du and Sie, Polish splits Pan and Pani, Thai marks politeness with ครับ and ค่ะ, Russian separates ты and вы. Every script has one of these decisions, and mixing the register reads wrong even when every word is pronounced perfectly.
• How stress is written. Russian usually omits stress marks but reads are ambiguous without them; Thai carries tone marks that change meaning; Polish applies a regular penultimate rule; German places compound stress automatically. Each guide names the stress lever for its language.
• Numbers and dates. Every language reads figures its own way — German numbers run back to front, Polish numerals take case forms, Thai years and prices need the grouping you want voiced. The rule is the same everywhere: write numbers the way they should be said.
• Punctuation is pacing. Commas and full stops are where the narrator breathes, in every language — and in Thai they are also the only word boundary marker the listener gets.

What stays the same

• One credit is about 1,000 characters in every language — the pricing page lays out the three packs.
• One voice, many languages — pick the voice once and keep it across every version; the voice library tags voices by fit.
• No subscription. Packs are one-time purchases and credits stay valid for 12 months.
• Existing videos can be re-voiced into any covered language with the original timing kept — see the AI dubbing guide.

Where to go next

If you are choosing your first language, start with the script checklist for that language above, then open the guides index — it groups every workflow on the site, from short-form video to audiobooks, by what you are trying to make.

Frequently asked questions

Do I need a different voice for each language?

No. Voices here are not locked to one language: one voice reads text in 30+ languages. Pick the voice that fits the brand, then write each version in its own language and keep the same voice across all of them — a bilingual or multilingual channel sounds like one channel.

How are non-Latin scripts handled?

Directly: paste Cyrillic, Arabic or Thai as written and the engine reads the script itself. Each language carries its own script decisions — stress marks in Russian, particles in Thai, register in German or Polish — and each language guide above covers the checklist for its script.

Does the price change with the language?

No. One credit covers about 1,000 characters in every language, so a five minute narration costs about four credits whether it is English, Arabic or Thai. Packs are one-time purchases, credits stay valid for 12 months, and there is no subscription.

My language is not listed here. Can I still use it?

Most likely yes. The library covers 30+ languages and the guide list grows over time; the workflow is the same for any of them — write the script in the language, let punctuation carry the pacing, check numbers and names, and proof-listen with a native speaker.

Can I dub an existing video instead of re-recording a script?

Yes. Re-voicing an existing edit while keeping its timing is the dubbing workflow, and it works across all 30+ languages with a single voice — the AI dubbing guide covers the process end to end.

Related

• Fish Audio voice library - how to audition and pick the voice behind every language.
• AI dubbing - re-voicing existing videos across 30+ languages with one voice.
• Fish Audio guides - the index of every workflow, tool and language on the site.