Spanish Text to Speech: Spain vs Latin American Voices
Spanish text to speech compared: Spain vs Latin American accents, seseo, voseo and vendor locale coverage as of October 2026, plus tips to pick the right voice.
Spanish text to speech works best when you choose the accent before the voice. Spain and Latin America differ in pronunciation and in forms of address such as tú, usted, vos and vosotros. As of October 2026, Azure lists 22 Spanish locales and Google's Chirp 3: HD covers Spain and US Spanish. Pick the audience first, then the voice.
Spain vs Latin America: what changes in the audio?
The differences are mostly in sounds and in the grammar of address. The table summarises the main points from Wikipedia's article on Spanish dialects and varieties. That article carries a banner saying it needs more citations, so treat its descriptions as general patterns, not exact shares of speakers.
| Feature | Spain | Latin America | |---|---|---| | c/z versus s (casa and caza) | Kept in northern Spain in all positions, and in south-central Spain at the start of a syllable | Not kept in the Americas, a pattern called seseo, where both sound the same | | Second-person plural | Vosotros, with its own verb forms, in northern and central Spain | Vosotros is not used. Ustedes covers formal and informal plural | | Informal singular "you" | Vos is absent | Vos is the prevalent informal form in Guatemala, El Salvador, Honduras, Nicaragua, Costa Rica, Paraguay, Argentina and Uruguay. It is absent from most of Mexico, Peru and the Caribbean islands | | ll and y | Yeísmo, where the two sounds merge, is found among most speakers | Yeísmo is found among most speakers. Around the Río de la Plata the merged sound is a sibilant, voiced or voiceless | | Final s | Weakened or dropped in southern Spain | Weakened or dropped in much of the Caribbean and in coastal and lowland areas |
These patterns matter for a voice because the text decides the form of address. A voice labelled Spanish will read "vosotros" or "vos" exactly as you write it, so the script must match the audience. No single accent suits every listener, so choose one on purpose.
How many Spanish voices do the major tools offer?
The counts below are as of October 2026. Vendors add and retire voices often, so check the current list before you buy or build around one.
| Vendor | Spanish coverage | Notes | |---|---|---| | Microsoft Azure AI Speech | 22 locales and 87 voices, including es-ES, es-MX, es-AR, es-CO and es-US | Most locales have two voices, one female and one male. Counts include Standard, multilingual, Dragon HD and MAI-Voice-2 rows | | Google Cloud Text-to-Speech (Chirp 3: HD) | es-ES and es-US, with the same 30 named voices | An es-MX Chirp 3: HD locale was not confirmed in the pages reviewed | | Amazon Polly | 3 locales and 12 voices: es-ES (6), es-MX (2) and es-US (4) | Voice engines differ by voice, including Generative, Neural and Long-form | | ElevenLabs | Spain and Mexico variants on Multilingual v2 and Flash v2.5. Spanish also appears in the v3 and v4 language lists | Per-model language coverage comes from the ElevenLabs models page | | Gemini API text-to-speech | Spanish listed by name; input language detected automatically | 30 prebuilt voices | | OpenAI text-to-speech | The page says voices are currently optimized for English | gpt-4o-mini-tts can be prompted for accent |
Azure has the widest count. A count alone does not settle the choice, though. A locale code tells you the target accent, not how a specific voice sounds, so listen to each candidate with your own script.
How should you write a Spanish script for your audience?
The locale does not set the form of address. Your text does, so decide these points before you generate audio:
- Informal "you": tú, vos or usted. Vos is the prevalent informal form in much of Central America and the Southern Cone, and it is absent from Spain.
- Plural "you": vosotros in northern and central Spain, and ustedes in Latin America, where vosotros is not used.
- Numbers and dates: Azure's say-as tags for cardinal numbers, ordinals, dates, currency and telephone numbers work for all Spanish locales, according to its SSML documentation. Its duration tag is supported only for English and Spanish.
- Punctuation: keep the opening ¿ and ¡ marks, because they show the sentence type to a reader and help the sentence read as written.
- Test one paragraph in two voices from different regions before you commit to either.
Mixing Spanish and English or brand names
Mixed-language text is a common request. The sources reviewed do not verify how a particular voice handles English inside a Spanish sentence, and OpenAI's own page says its voices are currently optimized for English. Test a sample that includes your brand names. If a name is misread, respell it phonetically in the script and listen again.
Spanish text to speech with OndaLabs
OndaLabs has a dedicated Spanish text to speech page. Its 30 prebuilt voices are usable in Spanish, delivery is directed in plain words with supported expressive tags, and the dialogue tool handles scripts with two speakers. For a consistent brand voice, custom voices are available from the Creator plan at $22 per month, which includes 3 custom voices. Custom voices can be cloned from a recording you have permission to use or designed from a text description.
The Free plan gives 2,000 credits per month, about 2 minutes on Onda Lite at 1 credit per character, so you can test a Spanish paragraph before paying. Starter at $9 adds Onda Flash, the most expressive model at 1.5 credits per character, plus commercial use. The verified feature list does not include regional accent selection, so the accent you hear comes from the voice and the script you write. Compare plans and prices before you choose.
For related reading, see the French text to speech guide and the Arabic text to speech guide.
Frequently asked questions
What is the difference between Spain Spanish and Latin American Spanish voices?
The main differences are pronunciation and forms of address. Most of the Americas use seseo, so c/z and s sound the same, while Spain keeps the distinction in much of the north. Spain uses vosotros for the plural "you", while Latin America uses ustedes, and vos appears in parts of Central and South America. Choose the accent and the form of address together, then check both in a test.
Is there a neutral Latin American Spanish voice?
The sources reviewed for this post do not confirm a single neutral Latin American voice at any vendor. Google's Chirp 3: HD lists es-US and es-ES, ElevenLabs lists Spain and Mexico variants, and Azure lists 22 national locales, as of October 2026. Listen to two or three candidates with your own script, then pick the one your audience expects.
Can Spanish text to speech switch between Spanish and English?
The sources do not verify code-switching quality for any specific voice. Gemini's TTS models detect the input language automatically, and Azure's multilingual voices speak the auto-detected language of the input text in the default locale. Test your own script, especially the brand names, before you publish.
Is there a free Spanish text to speech?
OndaLabs has a Free plan with 2,000 credits per month, about 2 minutes on Onda Lite, plus five saved clips. Google's Cloud pricing lists a free tier of 0 to 4 million characters per month for WaveNet and Standard voices. Amazon Polly lists 5 million characters per month for its Standard voices. Check each pricing page, because free tiers change.
How do I fix numbers, dates and abbreviations in Spanish?
Write them the way you want them spoken, and check the vendor's documentation for supported say-as tags. Azure's say-as tags for cardinal numbers, ordinals, dates, currency and telephone numbers work for all Spanish locales. Expand abbreviations in the script, then listen to the full output before you publish.