Unmanned Livestreaming

Born a Male—I Apologize

Whether it’s short-form video, long-form video, or livestreaming—as long as the medium is visual—women possess a natural advantage. Be it ASMR or dancing; fashion styling, makeup tutorials, or simply showcasing specific physical features—face, chest, waist, hands, buttocks, legs, or feet—any single body part can effortlessly generate massive traffic.

Even a mere silhouette, a voice, or a profile picture featuring a beautiful woman (even if the actual person behind the stream is male) is enough to constantly fuel the viewers’ imagination.

If you aim to acquire a significant amount of traffic, livestreaming is currently one of the best channels available. It offers real-time interaction, supports virtual gift-giving and monetary tips, enables affiliate marketing for product sales, and receives substantial promotional backing and support from major platforms.

Consequently, my current strategy involves “unmanned livestreaming.” I incorporate interactive gameplay mechanics into the stream—avoiding showing my face whenever possible, or at least minimizing direct human interaction—and delegate the rest to these interactive systems, allowing them to automatically respond to user input.

For instance, I might livestream the care of pets—fish, turtles, or cats. Users can trigger an automatic feeding sequence simply by sending a virtual gift (this allows the users to feel like they are “raising” the cat themselves, while sparing me the need for constant manual interaction). Similarly, I could livestream the care of potted plants, utilizing automated systems for watering, lighting, and monitoring temperature and humidity levels.

There are other variations as well—the most common being “sleep-interruption streams,” where users employ sounds, bursts of mist, or physical vibrations to wake the streamer or otherwise disrupt their sleep.

My plan for today is to host a “sleeping stream,” where viewers can send text commands—“1” to turn on the lights, or “2” to turn them off—to test how well this interactive concept works.

Home Assistant

The interactive mechanics powering my unmanned livestream setup are primarily built upon a combination of Home Assistant and ESPHome.

Here is a brief summary of today’s progress:

Previously, I used an Orange Pi 3B (2GB model) running Debian. On top of that, I installed Proxmox Virtual Environment (PVE), and subsequently deployed Home Assistant OS as a virtual machine within PVE.

A few days ago, I attempted to use the WireGuard add-on to establish a connection with a cloud-based WireGuard server. The connection itself was successful, but I was unable to configure domain-based proxy access.

Today, I tried using the Tailscale add-on instead. The connection was established successfully, yet I still could not access the Home Assistant interface via port 8123.

I had originally planned to redeploy my FRP (Fast Reverse Proxy) configuration today to enable domain-based access; however, I discovered that I had actually already configured domain access previously—the only issue was that my SSL certificate had expired.

Consequently, I simply renewed the SSL certificate. Once updated, I was able to access Home Assistant normally via its domain name—a workflow that is significantly more convenient.

Next, I noticed that the official add-on repository did not list a specific add-on for Xiaomi devices. To address this, I first installed the SSH add-on, then used SSH to execute the installation script for HACS (Home Assistant Community Store). Finally, I downloaded and installed the “Xiaomi Miot” integration directly through HACS.

[Continuing] in Xiaomi… I logged into my IoT account on Miot and added a Xiaomi Bluetooth light device. I could control its on/off state. Then I added the Xiaomi Smart Gateway 2, but it didn’t work; I couldn’t control it yet.

However, my goal is to do a live stream tonight while I’m sleeping. A YouTube live stream allows for interactive control of the lights. Once I can control the lights, I can configure MQTT and other settings later.

I’ve already set up the YouTube live stream to recognize user text and run local Node.js scripts. Tonight, I’ll try sending MQTT messages to control the lights, then start the live stream and test it again.

Blog Website

I also want to promote my blog website during the live stream.

I haven’t decided on a promotion method yet, and I don’t know if it’s feasible, or if it will affect traffic or get me banned.

For now, I’ll leave it at that and see how it goes.