Skip to content

ESPHome

Tip

See the Cloudcutter video guide for a complete tutorial on flashing with Cloudcutter and installing LibreTiny-ESPHome. Includes Home Assistant Add-On setup.

Because ESPHome does not natively support running on non-ESP chips, you need to use a fork of the project.

There are three basic ways to install and use LibreTiny-ESPHome. You can choose the option that best suits you:

  • ESPHome Dashboard (GUI) - for new users, might be an easy way to go; config management & compilation using web-based dashboard
  • command line (CLI) - for more experienced users; compilation using CLI commands, somewhat easier to troubleshoot
  • Home Assistant Add-On - using LibreTiny-ESPHome in Home Assistant, alongside your normal ESPHome installation - click the link for more info

Tip

You can use LibreTiny-ESPHome for ESP32/ESP8266 compilation as well - the forked version extends the base, and doesn't remove any existing features. Keep in mind that you might not have latest ESPHome updates until the fork gets updated (which usually happens at most every few weeks).

Find your device's board

Go to Boards & CPU list, find your board (chip model), click on it and remember the Board code. This will be used later, during config creation.

If your board isn't listed, use one of the Generic boards, depending on the chip type of your device.

Download ESPHome

For this, you need Docker, Docker Compose and Python installed. After running the commands, you'll have a running ESPHome Dashboard interface that you can connect to.

  1. Open a terminal/cmd.exe.
  2. Create a docker-compose.yml file in a directory of choice:

    docker-compose.yml
    version: "3"
    services:
      esphome:
        container_name: esphome-libretiny
        image: docker pull ghcr.io/libretiny-eu/libretiny-esphome-docker:latest
        volumes:
          - ./configs:/config:rw # (1)!
          - /etc/localtime:/etc/localtime:ro
        restart: always
        privileged: false
        network_mode: host
    
    1. You can change ./configs to another path, in which your ESPHome configs will be stored.
  3. Start the container using docker-compose up. You should be able to open the GUI on http://localhost:6052/.

Important

Read Getting started first - most importantly, the first part about installation.

It is very important that you have the latest version of LibreTiny installed (not libretiny-esphome - this is a different thing!) so that you don't face issues that are already resolved.

Assuming you have PlatformIO, git and Python installed:

  1. Open a terminal/cmd.exe, create esphome directory and cd into it.
  2. git clone https://github.com/libretiny-eu/libretiny-esphome
  3. cd into the newly created libretiny-esphome directory.
  4. Check if it works by typing python -m esphome

Tip

For Linux users (or if python -m esphome doesn't work for you):

  • uninstall ESPHome first: pip uninstall esphome
  • install the forked version: pip install -e .

Create your device config

  1. Open the GUI on http://localhost:6052/ (or a different IP address if you're running on a Pi).
  2. Go through the wizard steps:
    • New Device
    • Continue
    • enter name and WiFi details
    • choose LibreTiny
    • choose the board that you found before
    • select Skip
  3. A new config file will be added. Press Edit and proceed to the next section.
  1. Create a YAML config file for your device. You can either:
    • use python -m esphome wizard yourdevice.yml - type answers to the six questions the wizard asks, OR:
    • write a config file manually:
      yourdevice.yml
      esphome:
        name: yourdevice
      
      bk72xx:  # adjust accordingly: bk72xx or rtl87xx
        board: cb2s  # THIS IS YOUR BOARD CODE
        framework:
          version: latest
      
      logger:
      api:
        password: ""
      ota:
        password: ""
      
      wifi:
        ssid: "YourWiFiSSID"
        password: "SecretPa$$w0rd"
        ap:
          ssid: "Yourdevice Fallback Hotspot"
          password: "Dv2hZMGZRUvy"
      

Automatically generate config

Instead of adding components manually and writing everything from scratch, you can use UPK2ESPHome to generate a working config (for supported BK7231 devices only). If your device has a Cloudcutter profile, there's a high chance it can have a generated config.

Add components

Now, just like with standard ESPHome on ESP32/ESP8266, you need to add components for your device. Visit ESPHome homepage to learn about YAML configuration. If you want, you can upload an "empty" config first, and add actual components later.

Important

It's highly recommended to always include the web_server and captive_portal components - even in your first "empty" upload.

Adding these two components will safeguard you against accidentally soft-bricking the device, by e.g. entering invalid Wi-Fi credentials. The Web Server provides an easy way to flash a new image over-the-air, and the Captive Portal allows to easily open the Web Server on a fallback AP.

Build & upload

Close the config editor. Press the three dots icon and select Install. Choose Manual download and Modern format. The firmware will be compiled and a UF2 file will be downloaded automatically.

The command python -m esphome compile yourdevice.yml will compile ESPHome.

Now, refer to the flashing guide to learn how to upload ESPHome to your device. There's also info on using tuya-cloudcutter in that guide.

Advanced: LT configuration

Note

This part is for advanced users. You'll probably be fine with the default options.

All options from Options & config can be customized in the LibreTiny block:

yourdevice.yml
bk72xx:
  framework:
    version: latest
  lt_config:
    LT_LOG_HEAP: 1
    LT_UART_DEFAULT_PORT: 2
    LT_UART_SILENT_ALL: 0
(this is only an example)

Additionally, few options have their dedicated keys:

yourdevice.yml
bk72xx:
  framework:
    version: latest
  # verbose/trace/debug/info/warn/error/fatal
  loglevel: warn
  # suppress chip's SDK log messages
  # (same as LT_UART_SILENT_ALL above)
  sdk_silent: true
  # disable SWD/JTAG so that all GPIOs can be used
  # set to false if you want to attach a debugger
  gpio_recover: true
(these values are defaults)