A run and its result

One automation run, from queued to finished: what it cost, what it produced, and everything you can do with the result. This is where content is read, corrected, approved and sent onward.

/brands/{brand}/runs/{run}Anyone with access to the brand. A Member sees only the runs they started themselves.

What is on the page

The page is the same shape for every automation: a header with the status, a progress area while it works, four small cards of facts, the result itself, and the settings the run used. Only the result section changes from one automation to the next.

Part	What it tells you
Header	The automation's name, the generated title, and a status badge. Cancel Run sits here while a run can still be stopped, and Publish once a single post is finished.
Progress	Only while queued or running: a stage description, a percentage bar, and a row of stage icons filling in. Automations that write in pieces show a Live results card that fills in as each part is finished.
Started, Completed, Duration, Credits Used	Four cards. Credits are what this run actually cost, after any refund.
Results	The generated content, laid out for the kind of thing it is, with a toolbar above it.
Configuration Used	The settings this run was given, in plain words, so you can see what produced this output before running it again.

While it runs

You do not have to sit on the page. The run continues if you leave, and the header indicator in the app shows something is in flight.

A run can sit as Queued before it starts. Only one of your runs executes at a time, so anything else you start waits its turn rather than failing, and begins on its own when the current one finishes.

Cancel Run asks for confirmation, stops the run and refunds its credits. It is available while the run is pending, queued or running.

A run that gets stuck is cleaned up automatically after two hours and its credits are refunded.

What you can do with a result

The toolbar above the result is the same everywhere, with entries appearing only when they apply to that kind of output.

Action	What it does
Save	Puts the output in the brand's library under a title and optional tags, so it can be found and reused later.
Edit Design	Opens the visual editor on a post or slide, for moving, resizing and restyling the elements of the image. Shown only when the output has a design to edit.
Use In...	Hands this output to another automation as its input, for example sending a caption to the hashtag optimizer. Shown only when a compatible automation exists.
Copy and Export	Copy as text or JSON; export as JSON, CSV or plain text.
Schedule	Adds the content to the brand's calendar with a date, timezone, platforms and post type. A calendar's many posts are scheduled together.
Regenerate	Asks for a fresh version in place. Available where the automation supports it.

Publish Now in that menu is not built yet and stays disabled. Publishing a single post is done with the Publish button in the header instead.

Refining a result

When an automation leaves a resumable writing session behind, a Refine this result box appears under the output. Write what should change, in your own words, and it starts a new run that continues the same session rather than beginning from nothing.

That matters because a refined run keeps the context of the first one. Asking to shorten a caption or shift the tone does not mean re-entering every setting, and the result stays recognisably the same piece of work.

A refinement is its own run with its own credit cost, and it opens on its own page.

Images

A single post or carousel finishes with image prompts, not images. The Images panel under the result lists one card per slot, and nothing is generated, or charged, until you approve it.

On a card	What it is for
The prompt	Editable. Correct it before generating, and use Save prompt to keep the change.
Generate	Produces the image and charges its price, which is stated on the button. Generate all above the list does every pending slot at once.
Upload image	Use your own file instead of generating one. JPEG, PNG or WebP.
Skip	Leaves the slot blank.
Reject	Removes a generated image and asks what was wrong with it. The explanation is kept, so the brand learns from it rather than making the same image again.

After rejecting, Revise with Sparky opens the assistant beside this panel with your remarks already written out, so the prompt can be reworked in conversation and the result updates as it changes.

If a run produced no image at all, the header says why in plain words: image generation not configured for this app, turned off, no model set up, or not enough credits at the time. That line is the thing to read before assuming something broke.

A video post has no images. It carries one video prompt instead, and the clip appears only after the prompt is approved.

When a run fails

A failed run shows a red card explaining what happened, with Show technical details for the raw error where one was recorded. Its credits are refunded.

Run Again at the foot of a finished run offers two choices: repeat with exactly the same settings, or open the settings first and change something. After a failure, the second is usually what you want.

A run that fails because the brand's knowledge could not be read fails deliberately rather than inventing content. Generating against constraints a run never saw is worse than not generating at all.

Worth knowing

A Member can open only the runs they started. Editors and above see every run for the brand.

Some automations keep their results on their own screen. Opening a brand discovery run from the history takes you to the discovery review rather than here.

The header's Publish button appears only for a finished single post, and only for Editor, Manager and Owner. It posts to the accounts connected for that brand.

Scheduling a post whose images are still ungenerated warns you first, and offers to generate them before scheduling.