> For the complete documentation index, see [llms.txt](https://gaffa.dev/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://gaffa.dev/docs/features/browser-requests/actions/download-file.md).
# Download File
**Type**: `download_file`
Request a copy of the most recently viewed file in the browser.
Tips for using download_file action
* Gaffa keeps track of files the browser downloads, so you can run this action at any point afterwards to collect one.
* It works whether the file downloads automatically when the page loads or is triggered by an earlier action, such as a [`click`](/docs/features/browser-requests/actions/click.md).
* It returns the most recent download, so add a `download_file` action for each file if you're expecting several.
* Set a `timeout` long enough for the file to finish downloading. Larger files need longer, and the action waits until that limit is reached.
* Check your file type is supported: **.pdf, .jpg, .png, .gif, .bmp, .webp, .svg, .tiff, .tif** and **.img.**
* Each `download_file` action collects a file, which is then consumed. Add one action per file you're expecting.
* Don't add more `download_file` actions than there are files. Any extras will wait out their timeout and fail.
### Parameters
Name
Type
Required
Description
timeout
integer
false
The maximum amount of time the browser should wait for a file to download. Default: 5,000 (5s)
See [universal parameters](/docs/features/browser-requests/actions.md#universal-parameters).
### Files Supported
Currently, this only works with the following file formats: **.pdf, .jpg, .jpeg, .png, .gif, .bmp, .webp, .svg, .tiff, .tif, .img**
### Usage
#### Download a copy of a PDF and open it in the Browser
The following waits 20s for a file to download and then returns it.
```json
"actions": [
{
"type": "download_file",
"timeout": 20000
}
]
```
And the service responds with the file being in the action output:
```json
"actions": [
{
"id": "act_VHhrUbXjZSaYCPTqbBYD4acCzzeFGH",
"type": "download_file",
"query": "download_file?continue_on_fail=false&timeout=20000",
"timestamp": "2025-05-30T15:02:06.6615306Z",
"output": "https://storage.gaffa.dev/brq/downloads/5845df07-3749-424e-9c64-9602be19a857.pdf"
}
]
```
### FAQs
#### When do I use the download\_file action?
Use it when the thing you want is already a file — a PDF, an image, a report. It returns the original file rather than a version re-rendered from the page.
#### How do I download a PDF with Gaffa?
Point the browser request to the PDF's URL and add a `download_file` action with a `timeout` long enough for the file to finish downloading. The response gives you a storage URL.
#### Which file types does the download\_file action support?
It supports **.pdf, .jpg, .png, .gif, .bmp, .webp, .svg, .tiff, .tif** and **.img**. Other file types can't be downloaded with this action.
#### What timeout should I set for the download\_file action?
Set it to cover the time it takes to download the file. The 5,000ms default suits small images, while larger documents need more — 20,000ms is a good starting point.
#### How do I download a file that starts downloading on its own?
Just add a `download_file` action. Gaffa tracks downloads as they occur, so the file is collected whether it starts on page load or is triggered by an earlier action.
#### How do I download a file that's behind a link?
Add a [`click`](/docs/features/browser-requests/actions/click.md) for the link, then a `download_file` action with a long enough `timeout`. The click starts the download and `download_file` collects it.
#### How do I download multiple files in a request?
Add a `download_file` action for each file. Each action collects one download and removes it from the queue, so the next action picks up the following file. Give each a `custom_id` to tell them apart.
#### Why did my second download\_file action time out?
It was waiting for a download that hadn't happened. Each action consumes one file, so a second action only works once a second download has been triggered.
#### Does download\_file work on a PDF that opens in the browser?
Yes. It doesn't matter whether the PDF opens in the browser's viewer or downloads when you click a link. Either way, the action returns a storage URL for the file.
#### Where do I find the downloaded file?
The action's output field in the response contains a storage URL for the file, alongside the action's `id`, `type`, `query` and `timestamp`. The file keeps its original name.
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:
```
GET https://gaffa.dev/docs/features/browser-requests/actions/download-file.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.