Models

AssemblyAI offers several state-of-the-art speech recognition models, each optimized for different use cases, and a text-to-speech model that gives your application a voice. Choose the model that best fits your needs based on accuracy, latency, cost, and language requirements.

​

Text to Speech

​

Pre-recorded Speech to Text

​

Streaming Speech to Text

​

Add-on models

Add-on models enhance transcription accuracy for specialized domains. They work alongside your chosen speech model and are billed separately.

​

Medical Mode

Medical Mode (domain: "medical-v1") is an add-on that enhances transcription accuracy for medical terminology — including medication names, procedures, conditions, and dosages. It is optimized for medical entity recognition to correct terms that other models frequently get wrong. Supported models:

• Pre-recorded: Universal-3.5 Pro, Universal-2
• Streaming: Universal-3.6 Pro Realtime, Universal-3.5 Pro Realtime, Universal-Streaming English, Universal-Streaming Multilingual

Supported languages: English, Spanish, German, French Learn more: Medical Mode for pre-recorded audio | Medical Mode for streaming

​

Choosing the right model

​

Text to Speech

​

AssemblyAI-TTS-1

AssemblyAI-TTS-1 turns text into natural speech and streams the audio back as it is synthesized, so playback starts long before the sentence ends. It is built for voice agents: one WebSocket session stays open for the whole conversation, each reply is a request on it, and text can be sent as it arrives from an LLM. A plain HTTP API and an OpenAI-compatible endpoint cover one-shot requests and existing integrations. Pick from thousands of voices in the dashboard catalog or clone your own, steer pacing and tone through punctuation with the prompting guide, and get word boundaries to highlight text in sync with playback. US and EU data zones keep processing in the region you choose. Best for:

• Real-time voice agents that need the first audio of each reply as fast as possible
• Voice agents built on Universal-3.6 Pro Realtime, with Pipecat and LiveKit integrations
• Applications migrating from an OpenAI-compatible TTS endpoint
• Branded voices through voice cloning

Supported languages

Get started with AssemblyAI-TTS-1

​

Pre-recorded Speech to Text

​

Universal-3.5 Pro

Universal-3.5 Pro is our most powerful Voice AI model, designed to capture the “hard stuff” that traditional ASR models struggle with. It delivers state-of-the-art accuracy for entities, rare words, and domain-specific terminology out of the box, with code switching and optional prompting for more control. It’s also our fastest model, so you get the best accuracy without sacrificing speed. Best for:

• Applications requiring highest-accuracy transcription
• Medical scribes needing clinical grade transcription accuracy
• Sales intelligence / Call centers needing native code-switching
• Meeting notetakers / recruiting notetakers needing high-quality diarization

Supported languages

Try Universal-3.5 Pro here

​

Universal-2

Universal-2 offers accurate, cost-effective transcription across 99 languages with low latency. It supports code switching and optional keyterms prompting for domain-specific vocabulary (up to 200 words). Universal-2 is the go-to choice when you need reliable transcription across diverse languages. Best for:

• High accuracy at lower cost with broad language support
• High-volume, price-sensitive batch transcription
• Support for over 99 languages
• Recommended fallback when a requested language isn’t supported by Universal-3.5 Pro

Supported languages

Try Universal-2 here

​

Streaming Speech to Text

​

Universal-3.6 Pro Realtime

The most accurate model with the fastest word emissions for voice agents that demand the highest quality. Best-in-class accuracy with advanced prompting capabilities, including both keyterms prompting and native prompting. Supports 32 languages (see the list below). The previous flagship, Universal-3.5 Pro Realtime (universal-3-5-pro), is still supported for existing integrations, with the same features in 18 languages. Best for:

• Real-time voice agents
• Applications requiring premium accuracy
• Customer service voice agents needing elite entity accuracy
• IVR replacement / binary response detection in short utterances
• Agent assist and sales intelligence needing real-time speaker diarization, mid-session dynamic prompting
• Multilingual voice agents with native code-switching across 32 languages
• Compliance and verbatim recording — disfluency control via prompting

Supported languages

Learn more about Universal-3.6 Pro Realtime

​

Universal-Streaming Multilingual

A multilingual transcription model offering a good balance of speed and cost-effectiveness. Supports English, Spanish, German, French, Portuguese, and Italian. Features intelligent endpointing and keyterms prompting support for up to 100 words. Best for:

• Cost-effective real-time transcription across languages
• Cost-sensitive multilingual streaming across EN/ES/DE/FR/PT/IT

Supported languages

Learn more about Universal-Streaming Multilingual

​

Universal-Streaming English

An English transcription model offering a good balance of speed and cost-effectiveness. Features ~300ms word-by-word immutable transcripts, intelligent endpointing, and keyterms prompting support for up to 100 words. Best for:

• Cost-effective real-time transcription for English
• English-only real-time apps — fastest and cheapest streaming option for English

Supported languages

Learn more about Universal-Streaming English

​

Pricing

For detailed pricing information, visit our pricing page.

​

Text to Speech

Text-to-speech is billed per character of input text, not per second of audio.

​

Pre-recorded Speech to Text

​

Streaming Speech to Text

Streaming is billed per hour of session duration — the total time your WebSocket connection stays open — not per hour of audio sent. See Streaming Speech-to-Text billing for details. For volume discounts, please reach out to sales@assemblyai.com.

​

Next steps

• Explore Speech Understanding features like summarization, sentiment analysis, and more
• Learn about prompting: Universal-3.5 Pro prompting guide | Universal-3.6 Pro Realtime prompting guide
• Give your application a voice with the text-to-speech quickstart