customers
All customer stories
Top Voice AI companies are building with Assembly.
resources
Latest Release
Text-to-Speech API
Our first text-to-speech model, AssemblyAI-TTS-1, is built for voice agents, on live calls. It sounds natural, responds fast, and keeps callers on the line.
resources
Build voice agents that understand 32 languages with automatic language detection and real-time code-switching, and reply in English, Spanish, French, German, Italian, or Portuguese — no per-language pipeline required.
Language detection
LiveSpanish detected
"Buenos días, necesito ayuda con mi cuenta..."
French detected
"Bonjour, je voudrais vérifier mon solde..."
German detected
"Ich brauche Hilfe mit meiner Bestellung..."
Portuguese detected
"Olá, preciso cancelar minha assinatura..."
Global companies spend millions operating separate support queues per language — hiring bilingual agents, routing through IVR menus, and accepting degraded accuracy from English-centric ASR. When a customer code-switches mid-sentence, most systems break. Longer handle times, higher staffing costs, and a fragmented experience for non-English speakers all follow. AssemblyAI's multilingual pipeline eliminates the per-language bottleneck with a single model that detects, transcribes, and supports 32 languages — including real-time code-switching.
Languages from one realtime model with automatic detection.
Code-switching WER across five English↔European pairs — lowest of any streaming system tested.
Languages added in Universal-3.6 Pro Realtime, including Korean, Russian and Cantonese.
SLA with SOC 2 Type 2 certification.
Two ways to build
Run a full multilingual voice agent on AssemblyAI's speech-to-text, LLM Gateway, and Text-to-Speech, or drop industry-leading STT into the orchestration stack you already run.
Universal-3.6 Pro Realtime detects the caller's language, the LLM Gateway answers in it, and the Text-to-Speech API speaks it — on one API key and one bill.
Best for
Voice support in English, Spanish, French, German, Italian, and PortugueseAutomatic language detection across 32 languages on every turn4,000+ voices, with the language setting controlling the accentUS and EU processing at the same priceFree tier available · No credit card required
The multilingual STT layer for your voice agent. Works natively with LiveKit, Pipecat, Vapi, and Twilio — 32 languages, native code-switching across all 32.
Best for
Teams using LiveKit, Pipecat, Vapi, or Twilio as their orchestrator32 languages on Universal-3.6 Pro Realtime, all with native code-switchinglanguage_codes to pin a known language setReal-time language detection with confidence scoresKeyterm prompting for domain-specific vocabularyNo concurrency caps · Autoscaling included
Automatic language detection
The model identifies the dominant language in real time with confidence scores. Supports constrained detection with expected language lists — no IVR language menu required.
Multilingual transcription with code-switching
Universal-3.6 Pro Realtime natively handles code-switching across all 32 languages.
Language-aware LLM processing
Route transcripts through the LLM Gateway with the detected language code passed downstream, so the LLM responds in the caller's language without explicit prompting.
Localized voice response
The Text-to-Speech API replies in English, Spanish, French, German, Italian, or Portuguese, with the language setting controlling the accent and about 90 ms time to first audio.
Pipeline
Ingest multilingual audio
Detect language + transcribe
Route to language-aware LLM
Respond in caller's language
Universal-3.6 Pro Realtime + LiveKit — BYO stack
# LiveKit + AssemblyAI STT in a multilingual voice agent pipeline
from livekit.agents import Agent, AgentSession, TurnHandlingOptions
from livekit.plugins import assemblyai, cartesia, openai, silero
from livekit.plugins.turn_detector.multilingual import MultilingualModel
class MultilingualAgent(Agent):
def __init__(self):
super().__init__(
instructions=(
"You are a multilingual support agent. Respond in the "
"caller's detected language. Handle global customer "
"inquiries across all product lines."
),
)
async def entrypoint(ctx):
session = AgentSession(
stt=assemblyai.STT(
model="universal-3-6-pro",
language_detection=True, # include language metadata in turns
vad_threshold=0.3, # match Silero activation_threshold
),
llm=openai.LLM(model="gpt-4o"),
tts=cartesia.TTS(),
vad=silero.VAD.load(activation_threshold=0.3),
turn_handling=TurnHandlingOptions(
turn_detection=MultilingualModel(), # multilingual turn detection model
endpointing={"min_delay": 0.5, "max_delay": 3.0},
),
)
await session.start(room=ctx.room, agent=MultilingualAgent())
Universal-3.6 Pro Realtime handles 32 languages from one model with automatic language detection — one API call covers every market.
When callers switch languages mid-sentence — "I need help with mi cuenta, s'il vous plaît" — Universal-3.6 Pro Realtime transcribes each segment in the correct language without miscategorizing the call.
Universal-3.6 Pro Realtime goes beyond standard language codes with deep understanding of regional dialects — Quebecois French, Mexican Spanish, Brazilian Portuguese — capturing colloquial expressions and accent-specific pronunciation.
We require a leading edge speech-to-text provider that can meet our specialized needs: fast, accurate, targeted, and multilingual.
Super
.png)
Read more The transcription accuracy, reliability, and speed of AssemblyAI's API have greatly enhanced our operations.
Raj Shankar, SVP Product — Calabrio
Read more