River AI
A voice-controlled IoT AI Agent system built on n8n, the MCP protocol, and ESP32 hardware. The system supports voice-based reading of sensor data and control of multiple electrical circuits, using a low-code architecture to enable intelligent hardware interoperability.
Timeline (last 90 days)
River AI is an AI agent that can be commanded by voice; it receives instructions via speech-to-text, can read temperature, humidity, and distance data collected by an ESP32, and can switch 6 types of appliances on and off (lights, television, refrigerator, air conditioner, washing machine, and radio).
River AI is the graduation design project of Nabila Rahmadiani, an Information Technology student at Lampung University.
River AI's backend runs on a Linux virtual machine, containing n8n in Docker and a server.py service written in Flask, where server.py handles three tasks: saving sensor values, saving the on/off states of 6 devices, and receiving switch commands
River AI designed two independent workflows in n8n, serving respectively as an MCP Server and an AI Agent chat workflow, mapping to the dual-endpoint nature of MCP
River AI's chat model uses Google Gemini, and uses a Memory node with the Session ID sent from the webpage as the key to remember conversation context
All Facts (5)
River AI is an AI agent that can be commanded by voice; it receives instructions via speech-to-text, can read temperature, humidity, and distance data collected by an ESP32, and can switch 6 types of appliances on and off (lights, television, refrigerator, air conditioner, washing machine, and radio).
50%UnverifiedRiver AI is the graduation design project of Nabila Rahmadiani, an Information Technology student at Lampung University.
50%UnverifiedRiver AI's backend runs on a Linux virtual machine, containing n8n in Docker and a server.py service written in Flask, where server.py handles three tasks: saving sensor values, saving the on/off states of 6 devices, and receiving switch commands
50%UnverifiedRiver AI designed two independent workflows in n8n, serving respectively as an MCP Server and an AI Agent chat workflow, mapping to the dual-endpoint nature of MCP
50%UnverifiedRiver AI's chat model uses Google Gemini, and uses a Memory node with the Session ID sent from the webpage as the key to remember conversation context
50%