Automate post-capture steps with workflows

A workflow runs automatically when a matching capture is saved. It has three parts:

• a name
• a trigger that decides which captures it runs on
• a list of steps, each one a call to an MCP tool

Workflows run in the background, so saving a capture never waits for them. They don’t need an AI agent or an MCP client.

Open the Workflows window

Choose File → Workflows…. Existing workflows are listed on the left; select one to edit it.

To create one, click Add ▾ and pick Empty workflow or a template. Duplicate and Delete sit next to it.

Every workflow has an Enabled switch. Run now runs the selected workflow on the most recent gallery item and ignores the trigger, which is handy for testing without taking a new capture.

Triggers

A workflow runs when a saved capture matches all of its trigger settings:

Setting	Leave empty to
Capture modes, such as Slideshow	Match any mode
URL regex (optional), for example docs\.google\.com/presentation	Skip the URL filter
Title regex (optional), for example ^Q4 Plan	Skip the title filter

The URL and title come from the browser tab, so those filters only match captures that have one, like extension and slideshow captures.

Steps

Click Add step, give the step an id, choose a tool, and fill in Arguments (JSON). Most tools take an action, such as { "action": "tag" } for library_edit. Steps run top to bottom; reorder them with ▲ and ▼.

If a step fails, the run stops there, unless that step has Continue on error turned on.

Pick folder… inserts a folder path into the arguments for you.

Tokens

Text in {braces} inside the arguments is filled in when the workflow runs. Unknown tokens become an empty string.

Token	Value
{file_path}	Full path to the captured file. It updates after a rename step.
{file_name}, {file_stem}, {file_ext}	File name, without its extension, and the extension alone (no dot)
{title}	Browser tab title, for extension and slideshow captures
{url}	The page’s address, as it is
{url_host}, {url_path}	Host part of the URL, such as docs.google.com, and its path
{mode}	The capture mode, such as Slideshow
{date}, {time}, {datetime}	When the workflow runs; {date} is yyyy-MM-dd
{timestamp}	The same moment as Unix seconds
{n}	How many times this workflow has run
{temp_dir}	SuperCapture’s own temp folder, for files a later step uses

A later step can use an earlier step’s output as {step_id.key}, for example {export.output_path}. ~ at the start of a path means your home folder.

Templates

Template	What it does	Starts
Browser capture → rename with title	library_edit rename to {title}_{date} for captures with a URL	Enabled
Tag browser captures by URL host	library_edit tag with {url_host}	Enabled
Slideshow → export PDF to folder	export as PDF to ~/Slideshows/{title}_{date}.pdf	Disabled
Slideshow → social post (watermark + 1080×1080)	export as a 1080×1080 PNG, letterboxed and watermarked	Disabled
Any capture → export JPEG copy	export as JPEG to a path you set	Disabled
Any capture → dated archive folder	files copy into ~/SuperCapture Archive/{date}	Disabled
Any capture → copy to folder	files copy into ~/Captures	Disabled

Templates that write files start disabled. Set the output path or destination first, then turn on Enabled.

Example: file Google Slides captures by site

Set Capture modes to Slideshow and URL regex to docs\.google\.com. Then add two steps.

Step tag, tool library_edit:

{ "action": "tag", "ids": ["{file_path}"], "tags": ["{url_host}"] }

Step copy, tool files:

{ "action": "copy", "path": "{file_path}", "destination_dir": "~/Slides/{date}" }

Example: the social post template

The Slideshow → social post template is one export step. It stamps the slide with {url_host}, letterboxes it into a black 1080×1080 square and writes a PNG to the output_path you set:

{
  "path": "{file_path}",
  "output_path": "",
  "format": "png",
  "width": 1080,
  "height": 1080,
  "pad": true,
  "pad_color": "#000000",
  "watermark": "{url_host}",
  "watermark_at": "bottom_right",
  "watermark_opacity": 0.6
}

Workflows from earlier versions

SuperCapture 1.0.445 replaced the MCP server’s tools with fewer, broader ones. Workflows you saved before then are converted when the app loads them: each step moves to the tool that replaced it, with its arguments and the {step_id.key} tokens that read its result. A step that can’t be expressed the new way is kept but switched off, with a note saying why. The app backs up workflows.json before converting it.

Where workflows can write

The Approved Destinations list in Settings → AI MCP Info limits AI agents, not your workflows. A workflow can write wherever its arguments point, so check paths before enabling it.

What a workflow reads is limited the same way an agent’s is: a step’s source must be a library item or a file in SuperCapture’s temp folder, and files delete only removes files from the temp folder.

Workflows are saved in workflows.json in the app data folder.