Bulk Amazon Infographics
Research batches: a CSV or Seller Central export imports up to 300 ASINs, the dashboard submits them in automatic batches of 20 per call at 0.1 credit each, and four research jobs run in parallel. Image generation does not batch — there is no "generate for all selected" control, sets are started per product with two image jobs at a time, and the only unattended path is the public API at three concurrent runs per account. Colour variations can be repainted from a finished parent set; size, capacity and contents variations get no special handling.
Half of this page is a correction. Research in Graflio genuinely batches: import a Seller Central export and every ASIN in it can be scraped, cross-read against its category best sellers and turned into a brief for 0.1 credit each. Image generation does not batch. No control in the app renders sets for forty products at once, and the only unattended path is the API. Here is where that line sits.
What batches: research, at catalogue scale
The dashboard has a CSV / TXT import next to the ASIN box. It reads the file as text and pulls ASINs from any /dp/ link and from every bare ten-character token. A Seller Central export carries your own SKUs too, and those are often ten characters, so when real ASINs are present the importer keeps only tokens starting with B0. Duplicates are dropped, the file cap is 5 MB, and up to 300 ASINs land in the box.
Pasting or importing a long list just works. The endpoint accepts 20 items per call, so the dashboard splits the submission into batches of 20 for you and counts them off in the button — Adding 40/300… — until the list is through. Anything it could not queue is left in the composer rather than discarded, so a run that stops part-way is resumable. Two limits are real: the importer reads at most 300 ASINs from a file, and the shelf holds 200 products per account; if the shelf fills mid-run, the remaining ASINs stay in the box.
Behind a rotating residential proxy the research lane runs four jobs at once; without one, every request leaves from a single IP and walks into Amazon's captcha wall, so the lane drops to one. Each job is 14 steps: your listing, its category, the best-seller list, up to ten competitor listings, then one call that reads those reviews and writes the brief.
The arithmetic is the argument. A hundred ASINs is 10 credits, and the shelf holds 200 products per account, so researching all of them costs 20 credits — under a seventh of the 150 credits in the €49.99 plan. Rows are keyed by ASIN and marketplace, so resubmitting re-runs research on the same row, and one product on amazon.de and amazon.es is correctly two rows.
What does not batch: the image sets
Once the briefs exist, generation is per product: open a product, pick a style mode and a model, press generate, and the six or seven scripts in that brief render as one set. There is no multi-select on the shelf and no "generate for everything researched" button. The image lane runs two jobs at a time, so starting four sets means two render and two queue.
Cost per set is scripts times the model's price: 1 credit an image on Nano Banana 2, GPT-5 Image Mini or Grok Imagine, 2 on Nano Banana Pro (the default), 3 on GPT-5 Image and GPT-5.4 Image 2. A seven-image set on the default is 14 credits. Failed images are refunded automatically, as is a job lost to a restart. Sets download as a ZIP per product, up to 60 files; there is no catalogue-wide export.
The one unattended path: the API
To process a catalogue without sitting at the screen, that is what POST /v1/infographics is for. One call takes an ASIN plus a country, or a listing URL, and chains the two background jobs — research, then images — returning a run you poll. Six images is both default and maximum per run, three runs may be in flight per account, and wait: true holds the connection until the images land, ten minutes by default and fifteen at most. A real answer for forty SKUs, and a scripting job rather than a feature you click.
| Batches? | Limit | Cost | |
|---|---|---|---|
| ASIN import from CSV / TXT | Yes | 300 ASINs, 5 MB file | Free |
| Research jobs | Yes | Auto-batched in 20s, 4 running at once | 0.1 credit each |
| Image sets | No | Per product, 2 jobs at once | 1–3 credits per image |
| Colour variants | Within one product | 12 images × 8 colours per run | 1–3 credits per image |
| API runs | Yes, if you script it | 3 concurrent, 6 images per run | 0.1 + 6 images |
| Brand kits | No | One per account | Free |
Variation families are not handled as families
This is the claim most bulk pages in this category make and the one we cannot. Nothing in Graflio reads a parent ASIN, walks its children and builds a set per child. What exists is narrower.
Colour variants take images already generated for one product and re-issue them with the product repainted. Select up to 12 finished images, pick up to 8 colours, and each pair goes to an image model with one instruction: reproduce this exactly, change only the colour of the product. The colours come from the listing's own colour twister — scraped with the page, so each carries its variant ASIN and swatch, and the run fetches that variant's real photo as a reference — or you type them in with a photo of your own. Each output is priced as a generated image: 6 images across 4 colours costs 24 images' worth of credits.
So: for a colour family, build the parent set once and repaint. For a family varying by size, capacity, pack count, scent or contents, there is nothing here — each child is a separate row, a separate research job and a separate set.
Where bulk works
- Reading a whole catalogue: 200 ASINs for 20 credits.
- Triage before you spend a credit on pixels.
- Colour families, once the parent set is good.
Where it does not
- No multi-product image generation in the interface.
- Size, capacity and contents variations get nothing.
- One brand kit per account.
Keeping forty SKUs from looking machine-stamped
The usual failure of a bulk image tool is one template stamped onto everything, so a shopper who sees two of your listings sees the same picture twice. The defence is that the brief is per product: each SKU's pain points and USPs come from its own reviews and its own best sellers, so the argument differs even when the palette does not. Within a set, background treatments rotate — flood, macro, split, lifestyle, white — so no two images share a construction. The brand kit does the opposite job on purpose, overriding the researched style with your palette, fonts and logo so the catalogue reads as one brand.
Which approach fits your catalogue
Under about ten SKUs, ignore this page: research in one or two submissions and build the sets one at a time. Between ten and a hundred, research the whole catalogue first — it is the cheapest thing in the product — and generate only where a brief found something the current images do not say. Past a hundred, or on a recurring schedule, either budget the clicks honestly or put an engineer on the API. If you need one button turning a 400-SKU catalogue into 2,400 images unattended, we do not have it, and you should keep looking.
Common questions
Can Graflio generate images for my whole catalogue at once?
No. Research batches; image generation does not. Sets are started one product at a time, with two image jobs running concurrently. The only unattended route is the public API — three runs in flight per account, six images per run — which needs a client you write yourself.
How many ASINs can I import at once?
The CSV / TXT importer reads up to 300 ASINs from a file of up to 5 MB, including a Seller Central export, and drops SKU-shaped tokens when real B0 ASINs are present. You do not have to split it yourself: the dashboard submits the list in batches of 20 and shows its progress, leaving anything it could not queue in the composer. The shelf holds 200 products per account, so a full 300-ASIN import stops there.
What does researching a whole catalogue cost?
0.1 credit per ASIN: a hundred products is 10 credits, and the 200-product shelf limit is 20. The €49.99 plan carries 150 credits, and a new account starts with €10 of credits and no card. Image generation is where the money goes, at 1 to 3 credits per finished image.
Does Graflio handle Amazon variation families?
Only colour, and only as a repaint: up to 12 finished images from a parent re-issued in up to 8 colours, using the listing's own colour options and their reference photos. Nothing researches child ASINs, and size, capacity, pack-count and scent variations get no special treatment — each child is a separate product with its own research and its own set.
Can I keep separate brand kits for different clients?
Not in one account. A brand kit — palette, fonts, logo, default style mode — is stored once per account, so agencies running several brands swap the kit before each brand's runs or use separate accounts. Leaving it empty is also valid: each product's brief proposes its own palette.
This page is written by Graflio, so read our own column with appropriate suspicion. Competitor facts are the vendors’ published information as of August 2026 and change often — check before you buy.
Paste one ASIN. Get the whole set.
Graflio reads the live listing and its reviews, studies the best sellers around it, then renders a full set of on-brand infographics — every image still editable in a layer studio. €10 of credits free, no card.
Try Graflio free →