Home / FAQ
Questions & answers
Everything people ask before their first call
Short answers, in plain language. If something here is unclear, it is a bug in this page, not in you.
- ~1 secondTime from your text to spoken speech — measured at 0.9–1.3s in internal testing
- NothingWhat the other person needs — a normal phone call on a normal line
- NothingWhat you have to say out loud — you type, Dulce speaks
- A written recordWhat you keep afterwards — summary plus the full bilingual transcript
What is Hello Dulce?
Hello Dulce is an AI phone-call interpreter. You text what you want to say in your own language, Hello Dulce places the call and speaks it out loud in the other person’s language with a natural human-sounding voice, then texts every sentence of their reply back to you translated, while the call is still live. It works with any phone number and the other person needs no app, no account and no setup.
How does Hello Dulce work, step by step?
Four steps. 1) You open the app and write your message in your language. 2) You enter the number and Hello Dulce dials it. 3) When someone answers, Dulce introduces herself as calling on your behalf and speaks your message in their language; each sentence they say comes back to you as text in your language about a second later. 4) When the call ends you keep a written summary and the full bilingual transcript.
Do I have to speak during the call?
No. You type, Hello Dulce speaks. You can listen to the live call if you want to, but you never have to say a word in a language you are not comfortable with. Typing is also why the call stays accurate: what you wrote is exactly what gets said.
Does it sound like a robot?
No. Hello Dulce uses a natural, human-sounding voice with the local accent — Quebec French in Montreal, not textbook French — and it starts speaking about a second after you send your line, so the conversation keeps its normal rhythm instead of stalling. Most people on the other end simply answer as they would any other call.
How fast is the translation?
About one second. In internal testing on the Hello Dulce voice bridge, a typed message became spoken speech in the other language in roughly 0.9 to 1.3 seconds end to end. That is fast enough that the person on the phone does not feel a pause worth commenting on.
Which languages does Hello Dulce support?
Two different lists, and the difference matters. You can write to Dulce in Spanish, English, French, Portuguese, Arabic, Haitian Creole, Vietnamese or Mandarin — that is text, and the app is fully translated into all eight. The languages Dulce can currently SPEAK on the line are French (including Québécois), English, Spanish, German, Italian, Dutch and Japanese; Portuguese, Arabic, Mandarin, Haitian Creole and Vietnamese are shown in the app but marked as not yet speakable, because we would rather tell you than have you find out on a call. The first market is Spanish and Arabic speakers calling French-speaking institutions in Quebec.
Is Hello Dulce the same as a phone interpreter service?
The job is the same, the shape is different. A traditional over-the-phone interpreting service is a three-way call with a human interpreter, billed per minute — LanguageLine’s published pay-as-you-go rate is $3.95 per minute for audio. Hello Dulce has no third human on the line and no scheduling: you text, it speaks, and you get a written record. Human interpreters remain the right choice for legal proceedings, complex medical consent and anything where a certified interpreter is required.
Does the other person need to install anything?
No. They receive a normal phone call on a normal phone line. There is nothing to download, click or agree to on their side.
Does Hello Dulce tell people it is AI?
Yes. Dulce announces at the start of the call that she is a virtual interpreter calling on behalf of a named person. She never impersonates you, and she never invents an answer — if she is asked something you have not told her, she holds the line and asks you.
What does Hello Dulce cost?
Pricing is not published yet — Hello Dulce is in early access at bonjour.hellodulce.com. The design target is a cost per call far below per-minute human interpreting, because no third person has to join the line.
The four steps, again
Write what you want to say
In your language, the way you would actually say it. No perfect grammar, no script, no translation app in the other hand.
Dulce dials and speaks
A natural voice in their language — not a robot reading a translation. Dulce introduces herself as calling on your behalf.
You follow the call live
Every sentence comes back to you as text in your language, roughly a second after it is spoken. Send a new line any time and Dulce says it.
You keep the outcome
Dates, names, file numbers and next steps, written down in your language, with the full bilingual transcript attached.
Questions about specific calls
Does the school know they are talking to an AI?
Dulce introduces herself as calling on your behalf. She never pretends to be you, and she never invents facts — if the office asks something you have not told her, she politely holds the line and asks you.
Can I stay on the line and listen?
Yes. You hear the live call while reading the translation of every sentence, and you can send a new instruction at any moment.
Is my health information kept private?
Calls and transcripts belong to you. Dulce only says what you have written or approved, and you can delete a call and its transcript at any time.
What if the clinic puts the call on hold?
Dulce holds the line silently and pings you the moment a human comes back.
Can I use the transcript as proof?
Every call is saved with its timestamp, the number dialled and the full bilingual transcript, which you can export.
Will Dulce sit through the phone menu and the hold queue?
Yes — it presses the right options and waits. You get pinged when a person picks up.
Can Dulce give my personal details to the agent?
Only the details you have given it for that call. Anything else, it holds the line and asks you first.
Can I set this up for a parent who is not comfortable with apps?
Yes. Dulce works from plain text messages, and a family member can help write the first request or read the transcript afterwards.