Managed Agents / Bilingual answering service
#1 on EVA-BenchEVA-Bench is ServiceNow's independent benchmark for voice agents in enterprise workflows like customer support. It scores each agent on how accurately it completes the task and on the caller's experience.
#1 on Speech Agent ArenaSpeech Agent Arena is Artificial Analysis' independent benchmark for grading speech to speech models on enterprise tasks using human evaluators.
Bilingual answering service that answers in English and Spanish.
An AI answering service for small businesses that picks up in English or Spanish, takes the message in the caller's language, and pages your on-call tech when it can't wait. Call the sample line to hear the Spanish.
Trusted by leading teams
Ranked #1
on EVA-Bench and Speech Agent Arena
Answer in English or Spanish, then hand off in the same language
After hours, Spanish-speaking callers get an answer in Spanish without a Spanish-speaking night shift, and your team sees which language to call back in.
On your line, the greeting offers both languages, and a caller who answers in Spanish hears Spanish for the rest of the call, with no menu to press through. The sample line above plays only the Spanish side, so it greets in Spanish.
A caller who starts in English and finds Spanish easier can say so. With language detection on, the agent follows the switch and keeps the details already given.
The agent answers hours from what you provide and says so when it doesn't know. A caller who only needed the hours hangs up without leaving a message.
The saved request carries the caller's language, so the dispatcher who calls back in the morning can pick a Spanish speaker.
You decide what counts as urgent. The agent pages the technician and transfers the call, in whichever language the caller is using.
Spanish on the phone is not one thing
Callers in the US, Mexico and Spain expect different wording on recording and privacy notices. Check yours with a lawyer.
United States
Federal law allows recording when one party consents, but some states, California among them, require every party to agree. Tell callers in both languages that the call is recorded and that they are talking to an AI.
Mexico
A recorded voice is personal data. The federal data protection law requires a privacy notice, and it allows a short version in audio form when the data is collected by sound. Say it at the start of the call and point to the full notice.
Spain
A customer in Spain who calls a large company or a basic-service provider can ask for a person. Law 10/2025, in force since December 28, 2025 with 12 months to adapt, bars an automated channel as the only one, and the person must say who they are at the start.
Voices
Cartesia has Mexican and Castilian Spanish voices, and each speaks its own accent. Pick the one your callers speak, and test street names and phone numbers in full sentences before you publish.
This page is not legal advice, and the rules above change. Confirm the wording of your greeting with counsel in each place you take calls.
Built on best-in-class speech models
Cartesia handles listening, turn-taking and interruptions. You bring the instructions, the LLM and the tools.
- $0.03per minute
- #1on Speech Agent Arena and EVA-Bench
- 24/7answers every call, at any hour
Natural turn-taking
The agent lets callers finish, handles interruptions and pauses, and never talks over anyone.
Your choice of LLM and tools
Pick the LLM that is best suited for your agent's job, and give it tools to take action.
A voice that fits your brand
Pick from a library of natural, expressive voices so every call sounds like your company.
Check that the language travels with the message
This metric runs on every call and asks one question: did the request name the caller's language? A no is a Spanish speaker who may get an English callback.
Launch a bilingual answering service in three steps
Start buildingClone the template
Copy this use case into your Cartesia account and edit the instructions to match your policies.
Connect your systems
Connect the agent to the systems it needs, then give it a Cartesia phone number or bring your own.
Test the agent, then go live
Place test calls from the browser, read the transcripts, and score calls with your own metrics.
Enterprise-grade security
HIPAA compliant

SOC 2 Type 2

GDPR
Bilingual answering service FAQs
What is a bilingual answering service?
Bilingual answering services answer calls in two languages, here English and Spanish. This agent offers both at the start, replies in the one the caller picks, and saves the caller's language with the message.
How much does a bilingual answering service cost with Cartesia?
The agent bills $0.03 a minute, plus telephony and LLM usage, whether the caller speaks English or Spanish. The pricing page lists the telephony rate.
Can the agent switch languages in the middle of a call?
Yes, once you turn on language detection, a beta option in the agent's speech settings. The agent then listens for which of ten languages the caller is using and answers in that one: English, Spanish, French, German, Hindi, Russian, Portuguese, Japanese, Italian or Dutch. Turning it on moves speech recognition from Ink-2 to Deepgram Flux, so place a few test calls with your own pair of languages first. The sample line on this page stays in Spanish for the whole call.
Which languages can an answering service agent speak?
Cartesia voices speak 44 languages, and an agent speaks the language of the voice you pick. This page covers English and Spanish. For a line that answers in Spanish only, see the Spanish answering service. For other languages, see the multilingual AI phone agent.
How is an AI answering service different from a human bilingual answering service?
A human service puts a person on each call, and Spanish depends on who is on shift. This agent answers every call that reaches its number in English or Spanish and saves the caller's language with the message. It does only what your instructions and tools allow, so set it to page or transfer to a person for anything else. Call the sample line above and judge the Spanish yourself.
Does it sound the same in both languages?
It uses the voice you pick for the agent. Listen to your own script in each language before you publish, and have a fluent speaker check names, street names and numbers.
Where does the message go?
To the helpdesk, dispatch tool or shared inbox you connect with a webhook. Someone who can code sets that webhook up once, and Cartesia has no built-in connector to field-service software. The agent tells the caller the request is saved only after your system confirms it.
What happens to recordings and transcripts?
Your team can replay stored calls to check the Spanish. If you'd rather not keep them, an agent-wide setting stops the agent from storing transcripts, recordings and caller numbers. It applies to every call on that agent, and there is no per-call opt-out.
Does it speak Mexican Spanish or Castilian Spanish?
Both are available as voices. The agent speaks the accent of the voice you choose, so pick the one your callers use and listen to your own script first.
Can it answer during business hours too?
Yes. It answers every call that reaches its number, so you can send it nights and weekends, calls your team can't get to during the day, or both. Forward calls from your existing business number, and set the on-call technician's number as the transfer target. Transfers work on Twilio or SIP trunk numbers.
Find your voice
Talk to an expert.
Connect with a member of our team and learn how Cartesia can help you build world-class voice experiences.
Start building.
Access our models via API and bring an agent into production with our SDKs and developer tools.
We’re hiring!
Build the future of interactive intelligence
Careers page



