How to use the Nature Remo local API for enhanced latency and reliability on Home Assistant

In this post, I'll delve into how to use the local RESTful API for reduced latency and enhanced reliability when controlling lights in Home Assistant with Nature Remo.

Background

IR-controlled lights are ubiquitous in Japan, even with the emergence of newer "smart" models that often rely on proprietary protocols and apps. Home Assistant offers various ways to control IR lights, with  Broadlink transmitters and the SmartIR integration being a popular choice. While this setup worked well for lights, controlling AC was less reliable due to the IR code complexity.

Commercial AC controllers, such as those from Sensibo and Nature Remo, offered superior AC control reliability. However, Sensibo's focus on climate control and the limitations of the Nature Remo integration at the time meant that controlling both lights and AC required two IR blasters per room—a redundancy I wanted to eliminate.

Nature Remo integration (cloud polling)

My initial attempt involved enhancing the Nature Remo integration. While another user had added light support, it only functioned with remote controls featuring a single "power" (toggle) button, which didn't suit my lights with discrete on/off buttons. Furthermore, the integration was outdated and incompatible with Home Assistant 2025.x.

My patch fixed the isues above, but the latency was noticeable, and I wasn't comfortable relying on the cloud for lighting control.

Setting up local control

Fortunately, Nature Remo offers a local RESTufl API on most of its devices, excluding the Nano model. This guide assumes your Nature Remo devices are already network-connected and have static IP addresses.

Here's simple schematic of this solution. This design allows IR code reuse across different Nature Remo devices and interfacing with the controllers as light entities in Home Assistant.

1. Learning remote codes

  • Aim the remote at the receiver and press a button; the IR receiver will store the last code it detected.
  • Use curl to retrieve the code and save it to a text file:
curl -H "X-Requested-With: XMLHttpRequest" http://<device IP>/messages
  • If using Windows, substitute curl with Invoke-WebRequest:
Invoke-WebRequest -Uri "http://<device IP>/messages" -Headers @{"X-Requested-With" = "XMLHttpRequest"} | Select-Object -Expand Content
  • Be sure to include the | Select-Object -Expand Context or Powershell will truncate the output

  • The output will be in JSON format, containing the frequency and data.
  • Repeat for each button you want to store (e.g., on, off, max).

2. Template Sensor Entities for IP Referencing

  • Create a template sensor in configuration.yaml for each Nature Remo device:
template:
  - sensor:
      - name: "nature_remo_local_bedroom"
        state: "Online"  # The state can be anything 
        attributes:
          ip_address: "<device 1 IP>"
  - sensor:
      - name: "nature_remo_local_office"
        state: "Online" 
        attributes:
          ip_address: "<device 2 IP>"
  • Setting the IP here allows you to pass the entity when calling REST commands, avoiding hardcoding commands per device.

3.  IR Codes as REST Commands

  • Create rest_commands for each saved button:
rest_command:
  nature_remo_on:
    url: "http://{{ state_attr(entity_id, 'ip_address') }}/messages" 
    method: post
    content_type: "application/json"
    headers:
      X-Requested-With: "XMLHttpRequest"
    payload: >
      {
        "format": "us",
        "freq": 37,
        "data":[3430,1782,389,481,387,479,...]
      }
  nature_remo_off:
    url: "http://{{ state_attr(entity_id, 'ip_address') }}/messages"
    method: post
    content_type: "application/json"
    headers:
      X-Requested-With: "XMLHttpRequest"
    payload: >
      {
        "format": "us",
        "freq": 37,
        "data": [3408,1808,397,469,363...]
      }

  • Note that the data in the payload is truncated in this example.
  • I'm only using on/off here, as using relative brightness and temperature gets even more complicated

4. Light Entities from Templates

  • Use light templates to create light entities that call the REST commands:
light:
  - platform: template
    lights:
      bedroom_light:
        friendly_name: "Bedroom Light"
        turn_on:
          - action: rest_command.nature_remo_on
            data:
              entity_id: nature_remo_local_bedroom
        turn_off:
          - action: rest_command.nature_remo_off
            data:
              entity_id: nature_remo_local_bedroom
      office_light:
        friendly_name: "Office Light"
        turn_on:
          - action: rest_command.nature_remo_on
            data:
              entity_id: nature_remo_local_office
        turn_off:
          - action: rest_command.nature_remo_off
            data:
              entity_id: nature_remo_local_office

Possible improvements

The local API has significantly reduced latency, and the system is now more resilient against internet outages. However, the setup and maintenance process remains somewhat complex.

Future considerations include developing a patch for the SmartIR integration or creating a new patch for the Nature Remo integration to streamline the onboarding process.




Comments

Post a Comment

Popular posts from this blog

Raspberry Pi-based Home Assistant Control Panel

Image
This is my take on a Home Assistant control panel platform. It's based on this guide to creating a Raspberry Pi minimal browser kiosk but with wake-on-motion and some tweaks to improve the experience. Background info, why I chose this over the other options, and plans for future expansion are at the bottom, for anyone interested. My plan is to install one in each room with room-specific widgets. Hardware RaspberryPi 3 B 8GB or larger high quality microSD Official  Rpi 7" touchscreen Official  Rpi camera v2 Smartipi touch2  case with camera mount USB keyboard for initial HA login Assemble the parts according to the included instructions but keep the microSD out for now for imaging. Raspbian lite setup Image Raspbian Lite to the SD using the imaging tool or other method. Create an empty file on the SD boot directory named ssh to enable SSH on boot. Create a file in the SD boot directory named wpa_supplicant.conf with the following content, replacing the variables with ...
Read more

KingNovyPC M6 N100 Alder Lake-N Slim Mini PC Review

Image
I recently purchased a KingNovyPC M6 N100 mini PC from  Aliexpress  to replace a laptop I carry between home and work. I chose the 16GB model with no storage for 27,651 JPY (~$183 US). I think this is a good deal even given the drawbacks below. Let me start with the bad first. Poor thermals The cooling is woefully inadequate for the N100 CPU. The CPU temperature frequently exceeded 90°C under moderately intensive activities, such as installing several Windows updates in parallel. The thermal throttling and lag became noticeable during these times. The fan speed shows "n/a" because there is no yellow fan sensor wire. Replacing the heatsink compound with high-quality silver paste* had negligible impact but removing the top cover dropped temps by ~10°C. That's still high but out of the danger zone. I can think of many ways KingNovy (or the ODM) could have improved this thermal design, such as using a metal case to increase dissipation, but that would have added to the weight...
Read more

Railway Museum (鉄道博物館, Tetsudō Hakubutsukan)

Image
If you're like me, you may have an image of trains being a legacy of the industrial revolution chugging freight across country but that utilitarian image is naive or at least incomplete.  Modern train and subway technology rivals aeronautics and the history of these marvels of engineering is almost romantic. The Railway Museum (鉄道博物館, Tetsudō Hakubutsukan)  has done an incredible job of showcasing the progress from the first steam locomotives in the late Edo period to the recent E6 shinkansen since it's opening in 2007. The museum is located in Omiya, Saitama the city North of Tokyo about an hour away from central Tokyo. From Omiya station there's a "new shuttle" line which takes you directly to the Museum station. British-built locomotive #1 from 1871. I wonder how hard it must have been to save all these trains from being turned into ships and planes for the war. In front of most exhibits are explanations in Japanese b...
Read more