Voice FAQ Assistant

A voice-enabled assistant that answers questions from a FAQ knowledge base using streaming agents and OpenAI models. Supports real-time transcription, reasoning, and tool-assisted responses.

Features

• Real-time speech-to-text (STT) using OpenAI audio API.
• Streaming agent responses with reasoning updates.
• Integration with a FAQ search tool to answer user queries.
• Maintains conversation history across multiple turns.
• Supports audio output for spoken responses.

References

https://platform.openai.com/docs/guides/voice-agents?voice-agent-architecture=speech-to-speech

https://openai.github.io/openai-agents-python/ref/voice/workflow/?utm_source=chatgpt.com#agents.voice.workflow.SingleAgentWorkflowCallbacks

https://developers.llamaindex.ai/python/framework/module_guides/loading/simpledirectoryreader/#_top

Installation

git clone https://github.com/<your-username>/voice-faq-assistant.git
cd voice-faq-assistant
pip install -r requirements.txt