AssemblyAI

AssemblyAI is an API-first Speech AI platform for developers and product teams building voice agents, AI notetakers, call analytics, medical transcription and dictation into their products.

Key capabilities include:

• Pre-recorded speech-to-text with Universal-3.5 Pro (its most accurate async model, 18 languages with code switching) and Universal-2 (99 languages), plus keyterm prompting, custom spelling, word-level timestamps and speaker diarization
• Real-time streaming transcription with Universal-3.6 Pro Realtime for voice agents and Universal-Streaming models, with automatically scaling concurrency
• Speech Understanding and guardrails for insights and safety, such as PII redaction, plus an LLM Gateway to apply language models to transcripts
• Text-to-Speech API with AssemblyAI-TTS-1, built for voice agents on live calls, and a Dictation API

Pricing: Freemium, pay-as-you-go. The free tier includes up to 185 hours of pre-recorded and 333 hours of streaming transcription with no credit card. Pre-recorded transcription starts at $0.15/hr (Universal-2) and $0.21/hr (Universal-3.5 Pro); text-to-speech is $15 per 1M characters. Volume and enterprise pricing are available.

Notable for its developer tooling (docs, API reference, cookbooks, playground), self-hosted and voice AI cloud options, and availability on the AWS Marketplace.