What is RUPA from Nagent?
RUPA is Nagent's AI image teammate. You describe the image and pick a preset if you need one, and RUPA renders a sheet of plates with your brand guidelines attached by default through Brand Lock. She then reads your brief back against your brand book, so the team sees exactly what she was told. She makes ad creatives, social images and product shoots, checks budgets before rendering, and files every image in a searchable Asset Library. RUPA sits on the Content team under CREA, the AI Content Chief of Staff.
The reason she exists is simple: an image model knows nothing about your brand unless it is told, every time. RUPA makes telling it the default. The full role is on the RUPA page.
What does RUPA make?
RUPA works from presets in seven groups: Freeform, Paid social, Organic social, Profile and brand, Product, Life-cycle and Editorial. A preset fixes the size and how many plates you get, so a Meta ad or a LinkedIn post comes out at the right shape. The default is No preset: any of three sizes (square, wide or tall) and one to four plates.
In practice that covers:
Ad creatives for paid social and display.
Social posts and carousels.
Product photoshoots: packshots on a studio sweep and lifestyle scenes, with your real product attached. See product photoshoots.
Brand and profile images.
Email and landing page visuals.
Editorial images for articles and reports.
Infographics (Beta): RUPA picks the numbers from your knowledge base and brief and notes where each came from. The studio sets every figure as text, so no image model draws a number, and a figure without a source is flagged "Source needed".
How does a RUPA render work?
A render in RUPA's studio takes five steps:
Pick a preset, or none. Your workspace can also save presets of its own, with instructions and up to eight references added to a built-in one.
Write the brief. Up to 4,000 characters. Make renders; Ask answers from what is already loaded, with no render and no cost.
Borrow a look, if you want one. A reference image comes with one of five instructions: match the style, reproduce the layout, change the headline, localise it, or remix it. Each says what it keeps and what it drops.
Attach what must appear. A product shot or logo goes in as an input image. A render takes up to ten. Pick a product from your catalog and its own images travel with the render, with the instruction to keep it exactly as shown.
Read the sheet. Plates arrive as a sheet. Each offers Ad set (one plate recomposed as square, tall and wide), Restyle, Relight and Cut out, and a Brand check card follows.
For a product shoot, type /shotlist and what the shoot is for. RUPA plans up to eight shots, each a packshot, lifestyle, flat lay, scene or mood image with its scene, camera angle and props, and each renders when a person presses its Render button.
What is Brand Lock?
Brand Lock is the reason RUPA's images look like your brand. It is on by default, and while it is on, your workspace's brand guidelines go with every render. The bench beside the composer shows the exact text that is sent, so nothing is hidden.
The guidelines come from your brand book in the Knowledge Hub. Nagent can draft it from your website, and a published guideline document you upload outranks anything read off the site. It holds your palette, typography, tone, recipes for each asset type and the patterns your brand never uses.
Your logo travels as a real image to be placed, never redrawn. When no logo travels, the model is told to draw no logo at all. You can switch Brand Lock off for one brief, and then no brand material goes with it.
Will RUPA render without asking?
That is the team's choice, and it changes as trust grows:
Before she is trusted, RUPA writes a plan and stops.
With approval per render, a plan card shows the estimated spend, and nothing renders until a person clicks "Approve and run".
Once trusted with standing approval, she renders, and every render is written to the audit log.
Whatever the level, her daily and monthly budget, a per-action cap for one batch and your workspace's own limits are checked before anything is sent. Over a limit, she refuses and names the limit.
How much does RUPA cost?
RUPA spends your workspace's Nagent credits, the same allowance every agent draws on; see pricing for plans. The estimate on a plan card grows with HD quality, the number of plates and each input image. For scale, generating a visual brand book from your logo costs about 40 credits, and as a rough guide from the pricing page a campaign build is about 40 credits and a rendered video about 700.
For large jobs, the /variants batch (Beta) runs one prompt over up to 1,000 rows of a CSV, such as forty products or twelve cities. RUPA quotes the batch, renders three rows as a preview, and running the rest waits for a person unless CREA is at L5 Lead. The run stops for a person once spend passes 110% of the quote.
Where do RUPA's images go?
Every saved render lands in the Asset Library, which is also the Assets tab of the Knowledge Hub. Each image is described by the model shortly after it lands, so the library can be searched by what a picture shows, not only by its file name. Admins see the whole library; other people see what they made or uploaded.
How does RUPA work with the rest of the team?
RUPA is part of Nagent, multiplayer AI for growth, so her work arrives from and returns to other teammates:
CREA leads the Content room and passes image requests to her.
HOOK, the ad creative strategy lead (Beta), writes production briefs a person can send to RUPA, and HOOK's GATE checks the plates that come back.
Alpha, the video teammate, starts every product shot from a product image in RUPA's library.
In a team room, typing @ and picking RUPA renders one brand-locked plate from the message, square, wide or tall, in standard or HD.
How is RUPA different from Photoroom, Canva or Pebblely?
Those are strong standalone tools. Photoroom is built for batch product editing with a Brand Library (Photoroom). Canva is a full design platform with Brand Kits and background removal (Canva pricing). Pebblely makes styled product scenes from one photo from $9 a month (Pebblely pricing).
RUPA differs in three ways. Your brand guidelines go with every render by default, and you can read the exact text sent. Spend is checked against budgets before anything renders, with a person approving where the team wants that. And she works inside the same team that writes the campaign, so a brief from HOOK or a request in the Content room reaches her without a hand-off between apps.
What does RUPA not do?
Her studio is in Beta. The header says so, and the Edit and Write modes are not built yet.
Three sizes only. Square, wide and tall; there is no 9:16 or custom crop yet.
The Brand check reads text, not pixels. It checks your brief for banned words and marks rules it cannot see as "Needs your eye". Look at every plate before you use it.
She does not make video. That is Alpha.
She does not publish. Images go to the Asset Library; posting to social media is not something Nagent does today, and MOXA, the social media lead, is coming soon.
She is not a bulk editor for an existing photo archive. For background removal across thousands of old photos, a dedicated tool fits better.
For more on governed image production, read what ecommerce marketers get wrong about AI brand governance.
Frequently Asked Questions
Does RUPA know our brand?
Only what your brand book says, sent with every render by Brand Lock, which is on by default. The bench shows the exact text sent, and your logo travels as an image to be placed rather than redrawn.
Does the Brand check look at the image?
No. It reads the brief against your brand book's banned words and marks rules a text check cannot see as "Needs your eye". It judges what RUPA was told, not what she made, so a person checks the plates.
Can RUPA keep our product exactly as it is?
Pick the product in Settings and its own images go with the render as references, with the instruction to keep it exactly as shown. Always zoom in on the label before using a plate.
How do I make AI generated images stay on brand with our colours, fonts and logo?
Put them in one brand book and attach it to every render rather than retyping them into prompts. In RUPA, Brand Lock does this by default and shows the exact text sent.
What is the difference between RUPA and Alpha?
RUPA makes still images; Alpha makes video. Alpha's product shots start from an image in RUPA's library, so the pack in a video matches the pack in your stills.
See RUPA render a brand-locked sheet for your brand: book a Nagent demo.