TEN Framework is an open-source framework for building real-time, multimodal conversational voice agents. A TEN agent is a declarative graph, a pipeline of nodes for RTC transport, STT, LLM, and TTS defined in a property.json file. This integration provides two extensions, one for each capability, so pointing an agent at Smallest AI means editing configuration rather than writing code. The smallest_asr_python extension handles what the caller says, streaming transcription from Pulse over WebSocket with word-level timing. The smallest_tts_python extension handles what the agent says back, streaming Lightning audio with low latency. Interruptions are handled by the framework: when the caller speaks over the agent, TEN flushes the TTS extension and Lightning synthesis is cancelled mid-stream.