Smart Devices Given Personalities, Chatted in Discord Experiment
A writer gave smart home devices personalities and let them chat in Discord, using n8n and hosted LLMs. The experiment pulled anonymized data from an ESP32…
SMART HOME DEVICES were given personalities and dropped into a discord chat in a hands-on experiment that used automation software, hosted language models and anonymized device data. The setup pulled information from an ESP32 soil moisture monitor, a smart video doorbell and an LG smart TV.
The result was less about utility and more about curiosity. The writer wanted to see what connected devices would say if they could speak for themselves.
How the discord experiment worked
The experiment used n8n, a self-hosted automation platform, to move data from the devices into an LLM. The writer said a local model was considered, but weak hardware made that impractical because smaller models were slow to generate responses.
To keep the chat moving, the setup used models hosted on Groq, including several with generous free limits. Each device was assigned a prompt that defined its personality, then fed anonymized data from Home Assistant.
That mattered for the design. Each device could only access its own data, which meant the only way it could learn what the others were doing was through the conversation in discord. The chat became the shared layer where the devices picked up context.
What data each device had
The writer said the automation pulled information from three devices: a soil moisture monitor built on an ESP32, a smart video doorbell and an LG smart TV. No device was given direct access to the others' feeds.
That separation shaped the exchange. The soil monitor stayed focused on its own readings, the doorbell on its own events and the TV on its own status, while the conversation tied the threads together.
There was a personal angle in the setup, too. The writer compared the habit of thanking chatbots such as ChatGPT with the way people usually treat smart home devices, which are expected to obey without much acknowledgment.
That contrast is the real hook here. Smart speakers, sensors and appliances usually disappear into the background, but a discord room gives them a voice, even if that voice is built from prompts and automated data feeds.
Why this matters for connected homes
The experiment shows how quickly ordinary household devices can be turned into something more social, or at least more performative, when automation and large language models are mixed together. It also shows the growing role of hosted AI services in making that kind of setup responsive enough to be interesting.
For users following smart home tools, the setup is a reminder that the line between device status updates and conversational interfaces is getting thinner. The devices still only knew what their own sensors or feeds provided, but discord became the stage where those fragments turned into a shared exchange.
The writer did not frame the project as a product demo. It was a controlled experiment, built to see how devices might behave if they were given voices, personalities and a place to talk. The clearest detail was simple: the chat drew from three devices, and the conversation stayed inside discord.
Source: news.sky.com
