To automatically download a file in Chrome, add **Download File**, enter the direct http(s) URL, configure **Repeat**, and choose **Save & Run**. Tab Automation waits for Chrome to finish the download, waits between cycles, and then runs the graph again.

This is a recurring interval, not a calendar schedule. Chrome must stay open, and each cycle starts only after the previous graph and its Repeat wait finish.

## Set up a recurring file download in Chrome

Use this version when the file has a stable direct URL, such as a report endpoint that returns a PDF, CSV, ZIP, or another file.

1. Open the page associated with the file and open Tab Automation from its toolbar action.
2. Click **New Automation**, then choose the open tab under **Select a Tab**.
3. Click **Add Step** and choose [Download File](/steps/download-file).
4. Paste the full http(s) file URL. A web page containing the link is not enough; Download File needs the link's direct destination.
5. Optionally enter a Downloads subfolder. Leave it empty to use the folder from Settings.
6. Click **Test download** once to confirm that the URL returns the file you expect.
7. Configure [Repeat](/steps/repeat). A new automation starts with a 30-second **Fixed** wait; you can choose **No wait**, **Fixed**, or **Random range**.
8. Choose **Save & Run**.

The wait belongs to the End block. It begins after the graph finishes, so a 60-second Fixed wait means “wait 60 seconds before the next cycle,” not “start precisely once per minute.” Repeat can also stop after a cycle count or a duration.

## Download a file whose link changes

Some portals keep the report page stable but generate a new file URL whenever the report updates. Save that live link before the download instead of hard-coding yesterday's destination.

1. Add **Refresh Page** if the source page must reload to expose the newest report.
2. Add [Wait / Delay](/steps/wait-delay) when the link appears only after the page finishes loading data.
3. Declare a name such as `reportUrl` under [Values](/steps/values).
4. Add [Save Value](/steps/save-value), pick the report link, and set **Save as** to its resolved `href`.
5. Add **Download File** and enter `{{reportUrl}}` as the URL.
6. Keep **Repeat** on to sample the link and download it again after each cycle wait.

Save Value reads one link you pick. It does not scan the page for every PDF, image, or archive. If the job is to visit a known list of pages first, the [link-walking guide](/blog/walk-through-links-in-chrome) explains how For Each Link uses the same assigned tab one URL at a time.

## Where recurring downloads are saved

Download File uses Chrome's [`chrome.downloads` API](https://developer.chrome.com/docs/extensions/reference/api/downloads) and writes beneath the browser's Downloads directory. Chrome extensions may suggest a relative path under Downloads, but they cannot use this step to write to an arbitrary absolute path.

The default subfolder is `Downloads / tab-automation`. You can change the shared folder in Settings or enter a folder override on the step. Nested folders may use letters, numbers, hyphens, underscores, and `/`.

When the URL ends with a safe filename, the extension keeps that name. Otherwise it creates a name from the URL host and the current timestamp. If the same name already exists, Chrome adds a counter instead of overwriting the earlier file.

The runner watches the download until Chrome reports it complete. An interrupted download or one that has not completed within 60 seconds records a warning for that step. The file itself is not copied into Activity or a saved Value.

## Chrome automatic downloads are a different setting

Chrome's **Automatic downloads** site setting controls whether a website may download several related files. It does not create a recurring job, remember a file URL, or start a new download after a wait. Google's [Chrome download guide](https://support.google.com/chrome/answer/95759?hl=en) also covers the default download location, the “Ask where to save each file” preference, and PDF handling.

Tab Automation initiates the URL you configured when its graph reaches Download File. Chrome still owns the resulting download and its security checks. Do not dismiss a dangerous-download warning merely because the request came from an automation you created.

## Limits to know before you start

- **Direct http(s) URLs only.** Download File does not accept `blob:`, `data:`, local disk, or browser-internal URLs.
- **No calendar scheduling.** Repeat provides a wait between cycles, up to 3600 seconds where allowed. It cannot run “every day at 9:00.”
- **Chrome must stay open.** The automation does not run in the cloud or continue after you quit Chrome. Running automations reschedule when Chrome opens again.
- **Not a bulk download manager.** The step does not scan a page, filter by file type, create a parallel queue, or discover every downloadable link.
- **No custom request builder.** The step downloads an http(s) URL. It does not let you set custom headers, submit a POST body, or parse the response into [Values](/steps/values).
- **Downloads directory only.** A folder setting is a subfolder under Chrome Downloads, not a full path elsewhere on the computer.
- **Chrome protections still apply.** Browser and operating-system security controls may block or interrupt unsafe files.

Use the known-URL recipe when the same endpoint produces the file each time. Use Save Value first when the page publishes a fresh link. In both cases, Repeat is an interval between completed cycles—not a promise that the file will arrive at an exact clock time.