Production Automation is live: automate the shoots you keep repeating in Studio

If you submit the same brief in Studio for every new product, the same art direction, the same model, the same crops, with only the garment changing, you can now turn that brief into an automation. Production Automation is live for every Uwear workspace, as the Automation tab next to Studio: wire the settings you reuse to a trigger once, publish, switch it on, and every new upload, tag, or outfit runs the same shoot on its own.
What shipped
- -A node canvas for production: triggers, asset cards, shoots, edits, video, a quality check, and export, wired together like a production line.
- -Three triggers: when you upload clothing, when a tag is added to a product, or when an outfit is created. You can also start an automation by hand, or from your own systems through the API.
- -Versioned publishing: a published version is frozen for good; edits start a new draft while the live version keeps running.
- -Cost before commitment: the canvas totals credits per run and the worst case with re-shoots before you publish, and every run is recorded with what it made and what it spent.
A real automation: sunglasses uploaded, two photos 39 seconds later
The automation in the image above is live in our own workspace. Its trigger listens for clothing that finishes processing. Its Shoot card is wired to a saved art direction called Tropical Pop and to the Gemini Flash 2 Lite engine, and a Camera Edit card behind it re-crops whatever the shoot produces.
We uploaded a pair of sunglasses and did nothing else. The trigger fired on its own, the automation finished in 39 seconds, all 5 steps succeeded, and it spent 4 credits. These are the two results it created:


That is the whole idea. The brief you used to submit by hand for every new product now starts itself, carries the art direction you saved into every shot, and lands in your library ready to review.
What starts an automation
Three things you already do in your workspace. When you upload clothing, the automation starts as the upload finishes processing. When a tag is added to a product, it starts for the tags you chose to watch. When an outfit is created, the look gets its photos. In the builder these read as When clothing is ready, When a tag is added, and When an outfit is made. You can also press go by hand on a draft, or start an automation from your own systems through the API.
Here is the tag trigger at work. A second live automation in our workspace watches one tag. We tagged a graphic tee, and the shoot came back with a jump by the pool:


The three automations to build first
Every store we talk to repeats the same three shoots. Each is a handful of cards from the library, wired and published in the canvas. The photos below are real Uwear outputs of the garments shown, generated with Gemini Flash 2.
Product page set on upload
When you upload clothing, Shoot, Camera Edit, Check quality, Export tagged pdp-ready. Every new garment gets its listing photos without a brief.



Campaign tag to lifestyle set
When a tag is added, Shoot in the campaign art direction, Video, Export to your DAM by webhook. Tag the products in the campaign and the campaign look shoots itself, like the tee above.
New outfit to Shop the Look
When an outfit is created, Shoot, Swap Model, Check quality, Export tagged shop-the-look. Merchandise a look and it gets shot on two models.


Which steps an automation can chain
The node library covers the production line end to end. Shoot cards make new photos of whatever started the automation, paint them from a prompt, or bring a frame to life as video. Edit cards put the same clothing on a different model, reframe a shot, or apply a described change. Asset cards wire in the model, the location, the art direction, the prompt, and the AI engine, so swapping one card updates every step that uses it. A quality check scores the takes and separates the good ones. Export tags the finished photos or sends them to another system by webhook. The full library is on the Production Automation page.
Quality check: a step you place, not a review we force
Place a Check quality step anywhere in the chain and the automatic reviewer scores every take, keeps the good ones, and re-shoots rejects up to the allowance you set. Each re-shoot is billed like the original take, and the worst case is totaled before you publish. Leave the card out and photos flow straight through.
What an automation costs, and who is in control
An automation spends credits at the same rates as running the steps yourself, and the rules are strict because nobody is watching. The canvas shows the estimated credits per run. Each run reserves its estimate when it starts; if credits are short, it is rejected up front instead of stopping halfway. Steps that fail on our side refund what they reserved, and the run record says so plainly.
Nothing fires by surprise either. Uploading clothing or applying a tag first shows which automations will fire, how many runs that means, and up to how many credits, before you commit. Every automation has its own on and off switch, and one master switch stops the whole company; stopping records a reason in the audit trail, and runs already in flight finish under their own retry rules.
Published versions are frozen. Editing a live automation starts a new draft while the live version keeps running, and run history names the exact version behind every photo.
Where it fits beside Batch Workflows and MCP
Batch Workflows remain the way to run one large job you start yourself: a whole season submitted at once. Uwear MCP remains the way to drive production from the AI assistant you already use, with a costed brief you approve in chat. Production Automation covers the third case: the shoot nobody should have to start, because the trigger already knows. All three share the same library, the same art direction, and the same credits.
Build your first automation in the Automation tab
Open the Automation tab, wire a trigger to the settings you reuse, publish v1, and switch it on. Your next upload does the rest.
Prefer one big run you start yourself? See Batch Workflows.