Upload your product data once. We fix what shopping channels reject, what marketplaces misread and what AI agents can't use, then hand back a clean file for each.
Your product data isn't read by shoppers first. It's read by shopping channels, by marketplaces, and now by AI agents, and each needs it a little differently. omniSKU gets it right for all three from one upload.
Google, Meta, Amazon, Etsy
Check every listing against fixed rules and approve or reject it.
Approved, or disapproved 02And the partners that connect you
Turn your data into their listing, and often into their shopping ads, exactly as sent.
Matched, or missed 03Shopping assistants and agents
Read your attributes to answer a buyer's question, and use what they can verify.
Readable, or unclearChannels reject listings that break their rules, and every rejected product is a product not selling. Fixing the data before you send it is what turns disapprovals into approvals.
Marketplaces, and the partners that connect you to them, turn your listings into shopping ads exactly as you sent them. Give them complete, accurate data and your products match more of what shoppers search for, so every listing and ad works harder.
Men's Organic Crew Tee, Navy
£22.00via a marketplaceNorthfield Men's Cotton Crew Tee, Navy
£19.50via a marketplaceClassic Crew Neck T-Shirt, Navy Blue
£15.99via a marketplaceHeavyweight Cotton Tee, Navy
£24.00Other storeEssential Navy T-Shirt, 3-Pack
£18.50Other storeClear titles, brand and colour match how shoppers actually search, so your product can appear when someone looks for it.
Marketplaces turn your listing into the ad, so stronger data means ads that earn their spend.
Improve your catalogue in omniSKU, then hand the cleaned file to your marketplace partner.
AI assistants and shopping agents answer a buyer's question from product data. When your attributes, identifiers and specs are complete and consistent, an agent can confirm your product fits the request, so it's considered instead of left out.
TrailPro Waterproof Walking Boot, Wide
£114.00Northfield Men's Leather Walking Boot
£109.00Ridgeway Hiker, Wide Fit, Waterproof
£99.00Material, fit, size, colour, compatibility: the details buyers ask for, filled in from your own data, so your product can match more requests.
Valid GTINs, brand and MPN, and variants grouped properly, so an agent can confirm exactly what you're selling.
Clear titles and descriptions, with units, sizes and colours written one way throughout, so nothing gets misread.
From a store export to a clean feed per channel, in three steps. Run it again for every catalogue you manage.
A CSV, XML or JSON file, or send it from your own system by API. Shopify, WooCommerce and Google feed columns are matched for you.
Rules fix formatting, AI fills gaps from your own data, and what only you know is left for you.
See every change and why. Edit anything, then take a feed per channel as a download, a hosted URL or an API call. Input and output are saved.
Send a CSV, XML or JSON file, or push straight from your own systems. Get clean feeds back in the format each channel reads, as a download, a hosted URL or an API call.
Authenticate with a token, send your products, and read the checked feeds back. Run it from a nightly job, your PIM or your store platform.
POST /api/v1/catalogs
Authorization: Bearer osk_…
{
"name": "Spring range",
"channels": ["google_shopping"],
"products": [
{ "sku": "TEE-1", "title": "Classic Cotton Tee",
"price": "19.50", "brand": "Acme" }
]
}{ "id": 42, "status": "scanning", "products_count": 1 }GET /api/v1/catalogs/42/channels/google_shopping/feed?format=xml
Authorization: Bearer osk_…<rss xmlns:g="http://base.google.com/ns/1.0" version="2.0">
<channel>
<item>
<g:id>TEE-1</g:id>
<title>Classic Cotton Tee</title>
<g:price>19.50 USD</g:price>
…POST https://your-app.example/hooks/omnisku
X-OmniSKU-Event: scan.completed
X-OmniSKU-Signature: sha256=9f2c…
{ "event": "scan.completed",
"data": { "catalog": { "id": 42, "status": "ready" } } }// Your endpoint replies with any 2xx. Verify the signature first.Google and Meta feeds, Atom and your own XML are read as they are. Columns are matched to each channel's fields for you.
Only the formats a channel can actually take are offered.
Connect omniSKU to an assistant that supports MCP, then just ask. It imports your catalogue, checks it, applies fixes and hands you the feed, and no credits are spent without your say-so.
create_catalog1,248 products importedscan_catalogconfirmed 1,248 creditsfix_catalogfixes applied and loggedgoogle_shopping.xmlDownloadAdd omniSKU's address to your assistant as a remote MCP server. Nothing to install.
Choose read-only or read and write. No token to copy, and you can disconnect from settings any time.
It works on your real catalogue and returns the same checked feeds as the app. Scans show their cost first, and there's no delete.
claude mcp add --transport http omnisku https://omnisku.com/mcpIt's built for people who work with product data. Here's who gets the most from it, and who doesn't yet.
Buy credits when you need them. One credit checks one product, once, on every channel you pick. Re-checks and fixes of products you've paid for are included (fair use applies to AI rewrites), and your first 50 SKUs are free.
A small catalogue or a trial run
Most single-store sellers
Multi-brand or agency catalogues
Marketplace-sized feeds
No. omniSKU works from a CSV, XML or JSON export. Nothing to install, no account access. If you'd rather automate it, the API takes the same data.
You can upload CSV, TSV, XML (Google and Meta feeds, Atom or your own structure) and JSON. Downloads depend on the channel: Google Shopping as XML, TSV (tab-delimited text) or JSON; Meta Catalog as XML, CSV, TSV or JSON; Amazon as CSV, TSV or JSON; Etsy as CSV or JSON.
Yes. omniSKU runs an MCP server, so an assistant that supports remote MCP connectors can upload a catalogue, check it, apply fixes and fetch your feeds for you. You sign in and approve access once, and you can make it read-only. Scans use your credits, so the assistant has to tell you the cost and get your agreement first. See the MCP docs.
Yes. Create a token in your settings, send products as JSON or a file, and read back issues and a cleaned feed per channel. Signed webhooks tell you when a scan or fix finishes. See the API reference.
Not directly yet. You get a checked, cleaned feed per channel. Download it and upload it to Google Merchant Center, Meta, Amazon or Etsy yourself, or give Google or Meta a hosted feed URL to fetch on a schedule.
AI assistants and shopping agents read product data to answer a buyer's question. We check that yours is complete, specific and consistent: attributes filled in, valid identifiers, clear titles and descriptions. That makes your products easier for an agent to read and compare.
We can't promise that, and nobody honestly can. Whether an assistant shows or recommends a product is its own decision. What omniSKU does is make sure your data gives it what it needs, so nothing is missing or unclear on your side.
Each product row in your file or API request. Variants on their own rows count separately. Re-checking the same products later is included. Fair use applies to AI rewrites, which are capped in proportion to what you've paid.
No. AI fixes use only information already in your product data. If a value can't be determined, it's left for you.
Yes. Your original data is kept and every change is logged per product, so you can see exactly what changed before you download.
See what channels would reject and what agents couldn't read. Your first 50 SKUs are free.
Or choose a file → No card · No store connection · Originals kept