While the configuration UI of Home Assistant is still in development, you can use this small webapp to modify your configuration. It's essentially an embedded Ace editor, which has syntax hightlighting and automatic linting for yaml files (and a ton of other features you can turn on and off). There is also an integrated file browser to select whatever file you want to edit. When you are done with editing the file, click the save-button (or hit CTRL+s/CMD+s) and it will replace the original file.
JT Martinez has done a wonderful job by implementing Material Design.
- Web-Based editor to modify your files with syntax highlighting and automatic yaml-linting
- Upload and download files
- Lists of available triggers, events, entities, conditions and services. Selected element gets inserted into the editor at the last cursor position.
- Home Assistant event observer (connect to HASS via WebSocket and see all the events that happen)
- Restart HASS directly with the click of a button
- SSL support
- Optional authentication and IP filtering for added security
- Direct links to Home Assistant documentation and icons
- Execute shell commands
- Stage and commit changes in Git repositories, create and switch between branches, push to SSH remotes
- Customizable editor settings (saved using localStorage)
If there is anything you want to have differently, feel free to fork and enhance. And if something is not working, create an issue here and I will have a look at it.
WARNING: This tool allows you to browse your filesystem and modify files. So be careful which files you edit, or you might break critical parts of your system.
There are no dependencies on Python modules that are not part of the standard library. And all the fancy JavaScript libraries are loaded from CDN (which means this does not work when you are offline).
- Copy configurator.py to your HASS configuration directory (e.g /home/homeassistant/.homeassistant)
- Make it executable (
sudo chmod 755 configurator.py
) - (Optional) Set the
GIT
variable in configurator.py toTrue
if GitPython is installed on your system - (Optional) Install pyotp if you want to use the time based
SESAME
feature (see below). - Execute it (
sudo ./configurator.py
) - To terminate the process do the usual
CTRL+C
, maybe once or twice
Near the top of the py-file you will find some global variables you can change to customize the configurator a little bit. If you are unfamiliar with Python: when setting variables of the type string, you have to write that within quotation marks. The default settings are fine for just checking this out quickly. With more customized setups you will have to change some settings though.
To keep your setting across updates it is also possible to save settings in an external file. In that case copy settings.conf whereever you like and append the full path to the file to the command when starting the configurator. E.g. sudo .configurator.py /home/homeassistant/.homeassistant/mysettings.conf
. This file is in JSON format. So make sure it has a valid syntax (you can set the editor to JSON to get syntax highlighting for the settings). The major difference to the settings in the py-file is, that None
becomes null
.
Another way of passing settings is by using environment variables. All settings passed via environment variables will overwrite the settings you have set in the settings.conf
file. This allows you to provide settings in you systemd service file or the way it is usually done with Docker. The names of the environment variables have to be named exactly like the regular ones, prepended with the prefix HC_
. You can customize this prefix in the settings.conf
by setting ENV_PREFIX
to something you like. ENV_PREFIX
can not be set via environment variable. For settings that are usually defined as lists (ALLOWED_NETWORKS
etc.) a comma is used as a separator for each value (e.g. HC_ALLOWED_NETWORKS="127.0.0.1,192.168.0.0/16"
).
The IP address the service is listening on. By default it is binding to 0.0.0.0
, which is every IPv4 interface on the system. When using ::
, all available IPv6- and IPv4-addresses will be used.
The port the service is listening on. By default it is using 3218, but you can change this if you need to. The former setting LISTENPORT
still works but is deprecated. Please change your settings accordingly.
It is possible to place configurator.py somewhere else. Set the BASEPATH
to something like "/home/homeassistant/.homeassistant"
, and no matter where you are running the configurator from, it will start serving files from there. This is needed if you plan on running the configurator with systemd.
Set ENFORCE_BASEPATH to True
to lock the configurator into the basepath and thereby prevent it from opening files outside of the BASEPATH
If you're using SSL, set the paths to your SSL files here. This is similar to the SSL setup you can do in HASS.
The configurator fetches some data from your running HASS instance. If the API isn't available through the default URL, modify this variable to fix this.
If you plan on using the restart button, you have to set your API password. Calling the restart service of HASS is prohibited without authentication.
Set IGNORE_SSL to True
to disable SSL verification when connecting to the Home Assistant API (while fetching entities etc., not in your browser). This is useful if Home Assistant is configured with SSL, but the configurator accesses it via IP, in which case SSL verification will fail.
If you want to enable HTTP basic authentication you can set the desired username here. The :
character is not allowed.
Set the password that should be used for authentication. Only if USERNAME
and PASSWORD
are set authentication will be enabled. You may provide the password as a SHA256-hash with the prefix {sha256}
. For example PASSWORD = "test"
is functionally equal to PASSWORD = "{sha256}9f86d081884c7d659a2feaa0c55ad015a3bf4f1b2b0b822cd15d6c15b0f00a08"
. The hash will be converted to lower case automatically. Using the hash provides extra security by not exposing the actual password in plaintext in your configuration.
The credentials in the form of "username:password"
are now deprecated and should be removed from you configuration. Replace it by specifying USERNAME
and PASSWORD
. It will still work though to ensure backwards compatibility.
Limit access to the configurator by adding allowed IP addresses / networks to the list, e.g ALLOWED_NETWORKS = ["192.168.0.0/24", "172.16.47.23"]
. If you are using the hass.io addon of the configurator, add the docker-network 172.30.0.0/16
to this list.
List of statically banned IP addresses, e.g. BANNED_IPS = ["1.1.1.1", "2.2.2.2"]
Ban IPs after n failed login attempts. Restart service to reset banning. The default of 0
disables this feature. CREDENTIALS
has to be set for this to work.
Files and folders to ignore in the UI, e.g. IGNORE_PATTERN = [".*", "*.log", "__pycache__"]
Set this variable to True
to enable Git integration. This feature requires GitPython
to be installed on the system that is running the configurator.
To push local commits to a remote repository, you have to add the remote manually: git remote add origin ssh://somehost:/user/repo.git
Verify, that the user that is running the configurator is allowed to push without any interaction (by using SSH PubKey authentication for example).
If set to true
, directories will be displayed at the top.
If set to somesecretkeynobodycanguess, you can browse to https://your.configurator:3218/somesecretkeynobodycanguess
from any IP, and it will be removed from the BANNED_IPS
list (in case it has been banned before) and added to the ALLOWED_NETWORKS
list. Once the request has been processed you will automatically be redirected to the configurator. Think of this as dynamically allowing access from untrusted IPs by providing a secret key (open sesame!). Keep in mind, that once the IP has been added, you will either have to restart the configurator or manually remove the IP through the Network status to revoke access.
Instead of or additionally to the SESAME
token you may also specify a Base32 encoded string that serves as the token for time based OTP (one time password) IP whitelisting. It works like the regular SESAME
, but the request path that whitelists your IP changes every 30 seconds. You can add the SESAME_TOTP_SECRET
to most of the available OTP-Apps (Google Authenticator and alike) and just append the 6-digit number to the URI where your configurator is reachable. For this to work the pyotp module has to be installed.
HTTP requests include the hostname to which the request has been made. To improve security you can set this parameter to yourdomain.example.com
. This will check if the hostname within the request matches the one you are expecting. If it does not match, a 403 Forbidden
response will be sent. As a result attackers that scan your IP address won't be able to connect unless they know the correct hostname. Be careful with this option though, because it prohibits you from accessing the configurator directly via IP.
To modify the default prefix for settings passed as environment variables (HC_
) change this setting to another value that meets your demands.
Define a notification service from your Home Assistant setup that should be used to send notifications, e.g. notify.mytelegram
. The default is persistent_notification.create
. Do NOT change the value of the NOTIFY_SERVICE_DEFAULT
variable! You will be notified if your HASS_API_PASSWORD
, SESAME
or PASSWORD
password seems insecure. Additionally a notification with the accessing IP will be sent every time the SESAME
token has been used for whitelisting. To disable this feature set the value to False
.
Note regarding ALLOWED_NETWORKS
, BANNED_IPS
and BANLIMIT
:
The way this is implemented works in the following order:
- (Only if
CREDENTIALS
is set) Check credentials
- Failure: Retry
BANLIMIT
times, after that return error 420 (unless you try again without any authentication headers set, e.g. private tab of your browser) - Success: Continue
- Check if client IP address is in
BANNED_IPS
- Yes: Return error 420
- No: Continue
- Check if client IP address is in
ALLOWED_NETWORKS
- No: Return error 420
- Yes: Continue and display UI of configurator
Starting at version 0.2.5 you can add / remove IP addresses and networks from and to the ALLOWED_NETWORKS
and BANNED_IPS
lists at runtime. Keep in mind though, that these changes are not persistent and will be lost when the service is restarted. The API can be used through the UI in the Network status menu or by sending POST requests. A possible use case could be programmatically allowing access from your dynamic public IP, which can be required for some setups involving SSL.
api/allowed_networks
add
remove
curl -d "method=add&network=1.2.3.4" -X POST http://127.0.0.1:3218/api/allowed_networks
api/banned_ips
ban
unban
- Example:
curl -d "method=ban&ip=9.9.9.9" -X POST http://127.0.0.1:3218/api/banned_ips
HASS has the panel_iframe component. With this it is possible to embed the configurator directly into HASS, allowing you to modify your configuration through the HASS frontend.
An example configuration would look like this:
panel_iframe:
configurator:
title: Configurator
icon: mdi:wrench
url: http://123.123.132.132:3218
IMPORTANT: Be careful when setting up port forwarding to the configurator while embedding into HASS. If you don't restrict access by requiring authentication and / or blocking based on client IP addresses, your configuration will be exposed to the web!
Since the configurator script on its own is no service, you'll have to take some extra steps to keep it running. Here are three options (for Linux), but there are more, depending on your usecase.
- Simple fork into the background with the command
nohup sudo ./configurator.py &
- If your system is using systemd (that's usually what you'll find on a Raspberry PI), there's a template file you can use and then apply the same process to integrate it as mentioned in the HASS documentation. If you use this method you have to set the
BASEPATH
variable according to your environment. - If you have supervisor running on your system, hass-poc-configurator.supervisor would be an example configuration you could use to control the configurator.
- A tool called tmux, which should be pre-installed with recent AIO installers.
- A tool called screen. If it's not already installed on your system, you can do
sudo apt-get install screen
to get it. When it's installed, start a screen session by executingscreen
. Then navigate to your HASS directory and start the configurator like described above. Put the screen session into the background by pressingCTRL+A
and thenCTRL+D
. To resume the screen session, log in to your machine and executescreen -r
.
If you are using docker to run your homeassistant instance at home you can find corresponding docker images for the configurator on dockerhub. For usage visit the repository