AssemblyAI

Overview

AssemblyAI provides best-in-class speech recognition plus an intelligence layer: speaker diarization, sentiment analysis, auto chapters, content moderation, and LeMUR (LLM-powered Q&A on audio). Use it to turn audio/video files into structured, queryable data.

Setup

pip install assemblyai python-dotenv
export ASSEMBLYAI_API_KEY="your_api_key_here"

Core Concepts

• Transcript: The async job that converts audio → text. Submit a URL or file, poll for completion.
• Audio Intelligence: Optional enrichments added to the transcript request (diarization, sentiment, chapters, etc.).
• LeMUR: Apply LLMs to your transcript — summarize, answer questions, extract structured data.
• Real-time: Stream audio via WebSocket for live transcription.