AI Capabilities
AIUI includes core AI capabilities that allow your agent to understand users, speak responses, and perform comprehension and reasoning.
These three capability types are often combined to form a complete voice agent pipeline:
- Speech Recognition (ASR): Converts what the user says into text.
- Large Language Model (LLM): Understands context and generates replies.
- Speech Synthesis (TTS): Reads the generated result back to the user.
How To Choose
- If you want your app to "hear and understand" user input: use Speech Recognition
- If you want your app to "speak out" text: use Speech Synthesis
- If you want your app to handle Q&A, summarization, generation, or reasoning: use Large Language Model
Typical Flow
A complete voice agent interaction flow usually looks like this:
User speaks → Speech recognition → Text input → Large language model → Text output → Speech synthesis → User hears the reply
Continue Reading
- Speech Recognition: Learn how to capture user speech and receive recognition results.
- Speech Synthesis: Learn how to synthesize text into spoken audio.
- Large Language Model: Learn how to create model sessions, request replies, and read streaming output.