AssemblyAI

SaaS

Speech-to-text and voice agent APIs for transcribing recorded and live audio, with speech understanding, redaction and LLM access.

assemblyai.com
AssemblyAI homepage screenshot

About AssemblyAI

AssemblyAI provides Voice AI infrastructure through a set of APIs. Its documentation lists pre-recorded speech-to-text, streaming speech-to-text, synchronous speech-to-text, speech understanding, guardrails and LLM access, each aimed at a different part of a voice product.

The pre-recorded API transcribes audio and video asynchronously in 99 languages, which suits teams processing calls, podcasts or meetings in bulk. The streaming API handles live audio over WebSocket for voice agents and live captions, and the synchronous API returns a transcript in the same HTTP response for short clips, without polling or webhooks.

Beyond transcription, the platform adds speaker labels, summaries, sentiment and topic detection, inline redaction of personal data, content moderation and a gateway that routes requests to several large language models. Official JavaScript and Python SDKs are published on npm and PyPI, and a no-code playground lets people try the models first.

Key features

  • Asynchronous transcription in 99 languages
  • Live transcription over WebSocket
  • Single-request transcription for clips up to 20 minutes
  • Summaries, sentiment and topic detection
  • Inline PII redaction and content moderation
  • One endpoint routing to several LLMs with fallback
  • Dictation endpoint that formats text with an LLM

Good fit for

  • Call and meeting transcription
  • Live captions
  • Voice agents
  • Contact center agent assist
  • Podcast transcripts
Tags
speech to text
transcription
voice ai
api
voice agents
sdk

AssemblyAI: questions and answers

What is AssemblyAI used for?
AssemblyAI is a speech-to-text and voice agent APIs for transcribing recorded and live audio, with speech understanding, redaction and LLM access. It is a good fit for call and meeting transcription, live captions, and voice agents.
How much does AssemblyAI cost?
AssemblyAI is a paid product with no free plan. Listed pay-as-you-go rates are USD 4.5 for the Voice Agent API and USD 0.62 for the Dictation API, plus custom pricing.
Is AssemblyAI open source?
No. AssemblyAI is proprietary (closed-source) software. Open-source alternatives to AssemblyAI include LocalAI.
What are some alternatives to AssemblyAI?
AssemblyAI competes with Deepgram, Speechmatics and Rev. For open-source options, see Enlisted's ranked list of open-source AssemblyAI alternatives.

Open-source alternatives to AssemblyAI

See all

SaaS alternatives to AssemblyAI

See all