Have your greeting read aloud with ElevenLabs Eleven v3

Start from the listener, then press play

Start from the person who will listen, not from the voice you like best. The three samples above read the same line - "Happy birthday! I wish I could hug you today - this is the next best thing." - and each lands differently. Ava takes her time, Daniel says it plainly, Noa sounds as if the party has already started.

A letter to a parent wants warmth and time: Ava, or Mira, who sounds like somebody who has known them for years. A retirement or a thank-you after a hard year sits well with Daniel. Ari is grave, the one for goodbyes and the big anniversaries. For a birthday meant to be fun, try Noa or Yael; Tomer is easy and conversational, and Lena can hold a six-year-old through a long greeting.

The picker narrows it down by language, gender, age and style - warm, cheerful, calm, solemn, for a child, storytelling. Then press play. If the result sounds flat, move Stability down a little; if it sounds too theatrical for Grandpa, move it up.

Savta in Hebrew, the cousins in English, one voice

Eleven v3 covers Hebrew, and that matters here. A family in Haifa, Nazareth or Rishon LeZion gets a greeting in the language they speak at the table, not a translation. Mira can wish Savta a happy birthday in Hebrew and read the English card for the cousins in London in the same voice.

Write the greeting in the language it should be heard in, or ask the agent to write it that way - the voice reads what is on the page. Names are where any voice is most likely to trip: if Dedushka or Teta comes out wrong, spell it the way it sounds and make it again.

Accent is for English only. American, British, Australian and Indian change how an English greeting sounds; for Hebrew, Arabic and Russian it does nothing. Length is yours. The agent keeps a spoken greeting to 20-60 words, while Standard takes up to 2,500 characters - enough for a long letter to a parent far away. The price grows with the text and is on the button before you press.

When a voice alone is enough, and when it is not

Eleven v3 reads. It does not sing and it does not show a face, which makes the choice simple. When the words matter most and the listener has the phone to their ear - on a drive, in a hospital bed, at the kitchen table with their glasses in the other room - a voice greeting is the whole thing, and it costs less than a talking card.

When the face matters, make a talking card: the same voice, with Kling AI Avatar v2 moving the face and head of someone in a photo, or OmniHuman 1.5 moving the whole person. The grandmother on our talking-card page speaks with Ava's voice. For a song, Google Lyria 3.5 sings lyrics you approve first. For a video with voices in it, Veo 3.1 makes that sound in the same generation.