Bolna provides an end-to-end framework for creating voice-driven conversational applications by orchestrating multiple AI components including speech recognition, large language models, and text-to-speech synthesis. The platform supports configuration via JSON to quickly build voice agents that handle real-time conversations over websockets. It integrates with providers like Deepgram, Whisper, ElevenLabs, and supports various LLMs including OpenAI, DeepSeek, and Llama models.