provide a way for result variables to be displayed in a dashboard card

[hrabbach]

Describe the feature

I'd like to be able to run an automation and use the result variable in a dashboard card. Currently (unless the result is guaranteed to be less than 255 characters, in which case I could use an input text helper) this needs to be worked around with a trigger template sensor storing the result as an attribute, which is basically an automation, but needs to be configured in YAML mode with no way to easily manage and test.

Use cases

With the new AI tasks functionality, I can ask for a weather summary based on all my weather-related sensors, weather warnings from different services etc. - I'd like the result to be shown in a dashboard. The same would be true for other AI generated summaries about the states of multiple sensors, etc.

Current workarounds

Currently that had to be configured as a trigger template sensor, storing the result as a sensor attribute, and then displaying the attribute in a markdown card. While the automation can generate the summary, there is no way to pass the result anywhere that can be displayed in a dashboard card.

Anything else?

I'm not asking for a special dashboard card, although I guess that could be part of the solution

[karwosts]

Do you think there should be any memory (like remember the last run), or is this strictly an ephemeral: push button -> run script -> show result, and then that is gone when you reload the page?

I could maybe see the Script more-info page being one place that might be useful to show a response variable. Though that would be a JSON result and I'm not sure how we get from that to a nice display.

Or maybe a popup dialog result when running a script from the dashboard.

[hrabbach]

I see it as "display the result from the last run", that makes the most sense to me. Maybe with a "run actions now" option, but that's outside of the scope for me. My scenario is pretty simple:

1. automation runs on a trigger (in my case a timer)
2. automation result (>255 characters of text) shows up on the dashboard card
3. dashboard card content updates when the automation runs again

[karwosts]

I guess if we need to recall the result from a previous run (i.e. not attached to a run immediately requested in the frontend), than that data needs to be attached to some entity in the state machine somehow, as that's the only real way for frontend to get the data.

Maybe script/automation could have their last run response variable attached as an attribute... but I really think this is going against the current direction of development, trying to remove attributes and unnecessary data as possible from the state machine.

[Petro31]

Karwosts, it would likely be an action, like automation.get_variables. I don't see something like this getting implemented though.

Typically what I do is create a custom event that feeds a template entity for persisting values.

[alexruffell]

Before finally discovering this discussion, I had posted my own feature request. Mine focused on having a text entity with a much higher character limit. I do all of my automations in NR, and like the OP, I use AI to generate a custom forecast that can be announced as needed. The <60-word forecast is regularly updated and stored in a global variable in NR so it can be reused. I’d like to display it in HA’s UI, but as with the OP, the only workaround I found was to concatenate multiple text helpers or figure out how to use attributes.

I believe it would be very useful to have a simpler solution, such as a text entity with a much higher character limit. Allowing the display of "response variables" would help those using HA Automations, but those variables are not used in NR (Output Properties: msg.payload = results).

discussion-9004507

[Petro31]

using attributes isn't a workaround, that has a much larger limit and does what you want.

[hrabbach]

I guess a "result file" option in automations and a dashboard card that can display file contents would also be a possibility - I briefly looked into that, but the available cards don't seem to have a way to display multi line text files.

…

On Thu, 14 Aug 2025, 19:08 karwosts, ***@***.***> wrote: I guess if we need to recall the result from a previous run (i.e. not attached to a run immediately requested in the frontend), than that data needs to be attached to some entity in the state machine somehow, as that's the only real way for frontend to get the data. Maybe script/automation could have their last run response variable attached as an attribute... but I really think this is going against the current direction of development, trying to remove attributes and unnecessary data as possible from the state machine. — Reply to this email directly, view it on GitHub <#655 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/BLKKUMFYNBUH7CEMVCFTD4T3NS67VAVCNFSM6AAAAACD4AQMGGVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTIMJQG4YDCNA> . You are receiving this because you authored the thread.Message ID: <home-assistant/feature-requests/repo-discussions/655/comments/14107014@ github.com>

[alexruffell]

I looked at the markdown card to display the text but I think it still needs an entity, and no entities I know of support more than 256chrs. I read about creating a sensor and saving a time stamp as the main state and adding an attribute (not chr limited) for the longer text but I haven't figured that one out yet... plus it appears unnecessarily complicated to just store a longer string.

[evias]

Reviving a bit of an old thread because I saw your didn't have an answer here ...

# set_state.py
import requests
import json

def set_state(entity_id, data, host="localhost:8123", access_token=None):
    if access_token == None or len(access_token) == 0:
        raise Exception("Access token may not be empty.")

    response = requests.post(
        "http://" + host + "/api/states/" + entity_id,
        headers={
            "Authorization": "Bearer " + access_token,
            "Content-Type": "application/json",
        },
        data=json.dumps(data),
    )

    return response

Then just use like this:

# main.py
from set_state import set_state

if __name__ == '__main__':
  set_state("my_entity_name", data={
    "state": "OK",
    "attributes": {
      "friendly_name": "Don't forget attrs that already exist ;)",
      "custom_word": "Sincerely yours,",
      "and_another_one": "Simple as that!",
    })

And in your templates, instead of using state(), you would use state_attr("my_entity_name", "custom_word") for example.