← Node guide
Scripts

Run Script

Runs an Appium script you already have, on the phone this run is already holding. One line changes — the server address — and the rest of the file stays as it is.

script.run
The piece

Run an Appium script you already have, on the phone this run is already holding. Point it at $REMONODE_APPIUM_URL and the rest of the file stays as it is.

Reach for it when

  • You arrived with a working Appium suite and would rather not redraw it as thirty blocks to try this out.
  • One awkward stretch of a flow is already written in code, and the rest is easier as blocks.
  • You want the schedule, the alerts, the AI check and the multiple phones around a script you trust.
Run Script
Idle
In the bag

Every setting this block has. A field marked only when appears once you have chosen the mode above it — those are alternatives to each other, not extra things to fill in.

  • LanguageRequired
    starts at Python (Appium-Python-Client)

    The client library your script imports. Both are pre-installed, pinned — a script that passes today runs the same way in six months.

    Python (Appium-Python-Client)Node (WebdriverIO)
  • ScriptRequired
    starts at import os from appium import webdriver from appium.options.android import UiAutomator2Options # The one line that changes: the server comes from the run, not from a hard-coded address. driver = webdriver.Remote(os.environ["REMONODE_APPIUM_URL"], options=UiAutomator2Options()) try: print("elements on screen:", len(driver.find_elements("xpath", "//*"))) driver.save_screenshot(os.environ["REMONODE_ARTIFACTS"] + "/screen.png") finally: driver.quit() # ends this script, not the run

    Your script, as it is. Replace the Appium server address with os.environ["REMONODE_APPIUM_URL"] — capabilities are ignored, so leave them.

    only when Language = Python (Appium-Python-Client)

  • ScriptRequired
    starts at import { remote } from "webdriverio"; // The one line that changes: the server comes from the run, not from a hard-coded address. const url = new URL(process.env.REMONODE_APPIUM_URL); const driver = await remote({ protocol: "http", hostname: url.hostname, port: Number(url.port), path: url.pathname, capabilities: { platformName: "Android" }, }); console.log("page source:", (await driver.getPageSource()).length, "bytes"); await driver.saveScreenshot(`${process.env.REMONODE_ARTIFACTS}/screen.png`); await driver.deleteSession(); // ends this script, not the run

    Your script, as it is. Take the server address from process.env.REMONODE_APPIUM_URL — including its path, which carries this run's token.

    only when Language = Node (WebdriverIO)

  • Time limit
    starts at 300s

    How long the script may run before the step ends it. It is enforced from outside, so a script that hangs cannot hold the phone past this.

  • Memory
    starts at 512 MB

    Past this the script is stopped. Raise it only if a script really holds that much at once.

    256 MB512 MB1 GB
  • Environment
    starts at —

    One "NAME = value" per line, readable from the script. Values may reference earlier steps with {{ }}. Nothing else from the worker is visible — not our keys, not another run.

Build it

  1. Paste the script, then change one line

    Wherever it says http://localhost:4723 (or a Grid address), read the server from REMONODE_APPIUM_URL instead. Leave the capabilities block alone — the run's phone is already chosen, so it is ignored rather than obeyed.

    Run ScriptSettings
    Language
    Python (Appium-Python-Client)
    ScriptRequired
    driver = webdriver.Remote( os.environ["REMONODE_APPIUM_URL"], options=UiAutomator2Options(), )

  2. It starts where the last block left off

    The script is handed the session this run is already using — the build you installed, the account you logged in with, the screen the previous block left open. It is a step in the flow, not a fresh test.

    Launch App
    Run Script
    Run Script
    Analyze with AI
    Send Email

  3. Leave your teardown exactly as it is

    driver.quit() ends your script, not the run. The next block still has the phone. So do finally: blocks, pytest fixtures and afterEach hooks — nothing has to be deleted for the script to behave here.

    what the run does with your script's calls
    • webdriver.Remote(...)hands over the running session
    • find_element / tap / typegoes to the real phone
    • driver.quit()ends the script — the run keeps the device

  4. Pass values in, hand values back

    Environment takes NAME = value lines, and those may reference earlier blocks with {{ }}. Write JSON to REMONODE_OUTPUT and later blocks can read the fields off this step; save a screenshot into REMONODE_ARTIFACTS and it comes back with the run.

    the four names your script gets
    • REMONODE_APPIUM_URLthe server to connect to
    • REMONODE_OUTPUTwrite JSON here → the next block reads it
    • REMONODE_ARTIFACTSsave files here → they come back with the run
    • (yours)whatever you put in Environment

  5. Watch it the way you watch every other block

    Every command the script sends passes through the run, so the step reports what it actually did and how long each call took — not just a pile of stdout at the end.

    the step's report
    Script (python) finished in 6.2s
9 WebDriver commands:
  POST /session → 200 (session lent)
  GET  /session/:id/source → 200 150ms
  GET  /session/:id/screenshot → 200 1.2s

    plus anything your script printed

What comes out
  • Whatever your script wrote to REMONODE_OUTPUT, as fields later blocks can reference.
  • exitCode, how long it ran, and the list of files it left behind.
  • A screenshot, if the script saved one into REMONODE_ARTIFACTS.
Watch out
  • The server URL carries a path as well as a port. In WebdriverIO pass url.pathname as `path` — dropping it is the usual cause of a script that cannot connect.
  • The script runs sealed in: no access to the machine running it, no credentials of ours, and no way back to anything but the phone. Anything it needs has to arrive through Environment.
  • A script that hangs is stopped at the Time limit, and the step goes red. Set that limit to something you would actually wait for.
  • Only Appium-Python-Client / selenium / pytest / requests (Python) and webdriverio (Node) are installed. A script that imports your own helper package will not find it yet.
  • Running a device command (Android): pass the program and its arguments separately — {command: "settings", args: ["get", "system", "screen_off_timeout"]}. The `sh -c "one long string"` form looks the same and is not: a Pixel refuses it and a Galaxy runs the first word with no arguments.
  • It drives one device — the one this branch is holding. To run it on three phones, let the trigger fan out and this block runs once per phone.
Usually next to
Install AppLaunch / ActivateAI AnalyzeSend Email