Product data that Google and AI love.
feedlove fills the missing Merchant Center attributes from your store's own product pages: color, material, size, weight, product details, Q&A. Every value comes with a quote from the source, and the feed updates daily.
- Materialverifiedbamboo“Material: bamboo”
- Claimfiltered out“Antibacterial”Not on the product page
Online stores we work with
Wherever your customers search.
In Shopping filters, on product cards and in AI assistants' answers: they all work from what is in your feed.
This one matches what you asked for:
- 80% merino wool, 20% polyamide
- Adult size, 54–58 cm
- Hand wash at 30 °C
Illustration. How products are displayed is decided by Google and the AI providers.
The same product: 89 points instead of 23.
This feed isn't bad: it has a brand, a GTIN, a category and extra images. It still scores only 23, because it lacks everything shoppers and AI ask about. We fill that in from the product page, and show where every value comes from.
- Material, color, size, weight: for Shopping filters
- Highlights, product details, Q&A: for AI answers
- A detailed description and documents: for accurate answers
question_and_answerwith source quotesQIs it dishwasher safe?
ANo, hand wash only.
source: “Cleaning: hand wash only, not dishwasher safe.”
QIs it gentle on knives?
AYes, the bamboo surface is gentle on knife edges.
source: “The bamboo surface is gentle on knife edges.”
QHow big is it?
A38 × 28 cm, 1.8 cm thick.
source: “Size: 38 x 28 x 1.8 cm”
title- Bamboo cutting board 38×28 cm
price- €17.90
image_link- board-1.jpg
brand- KitchenPro
gtin- 5995800123453
additional_image_link- board-2.jpg · board-3.jpg
google_product_category- Kitchen & Dining > Cutting Boards
material- missingbamboo
color- missingnatural
size- missinglarge
product_length · width · height- missing38 cm · 28 cm · 1.8 cm
product_weight- missing950 g
product_highlight- missingGentle on knife edges · Groove catches juices · Hand wash
product_detail- missingDishwasher safe: no
question_and_answer- missingIs it dishwasher safe? – No, hand wash only.
structured_description- missingBamboo cutting board with a juice groove… (620 characters)
document_link- missinguser-manual.pdf
How AI-ready is your feed?
We run the system on a sample of your products, on your own product pages, and show you what's missing, what's risky and how much the score improves.
- AI readiness score now and after enrichment
- Missing fields, attribute by attribute
- Feed errors: price, stock and GTIN mismatches
- A real sample of your products, with source quotes
What it looks like in practice.
The source sits next to every value, and every decision is yours. Take a look inside.
Your feed has to answer what shoppers ask.
Google Shopping, AI Mode and ChatGPT shopping work from the attributes in your feed. If the feed is missing the material, the size or what the product is for, your product is left out of the answers, even when your store has all the details.
- Most store feeds only carry a title, price, image and description.
- The missing data is already on your product pages: in tables, descriptions and PDFs.
- We move it into your feed, verified and in the format Google requires.
Without these, the same product in your feed is just a title, a price and an image.
Three steps, and your feed improves on its own.
No plugin to install, and no need to replace your existing feed. Merchant Center gets a supplemental feed from us.
Share your feed URL
The same one Merchant Center uses today. A short questionnaire tells us your store's rules: tone of voice, words to avoid, what counts as an accessory.
We fill and verify
AI works from your product pages, and code traces every value back to its source. Prohibited claims are dropped, uncertain values wait for approval.
Merchant Center fetches it daily
A supplemental feed is published at a protected URL and refreshed every night. The client portal shows what changed and why.
| id | color | material | product_weight | product_highlight |
|---|---|---|---|---|
| CB-3828 | natural | bamboo | 950 g | Gentle on knife edges |
| HAT-001 | navy | merino wool/polyamide | – | Hand wash at 30 °C |
| SHF-2061 | orange/blue | steel | awaiting approval | 200 kg per shelf |
| TEE-BLK-M | black | 100% organic cotton | – | Unisex, crew neck |
from the store's own text
It runs on its own, every night.
Not a one-off project: the feed lives with your catalog. Costs under control, guardrails, approval and rollback.
Only what changed
Unchanged products are never regenerated: each night only new and changed products go through. That keeps it fast and cheap.
On failure, the last good feed stays
If a run fails, or would publish suspiciously few rows, we don't publish it: Merchant Center keeps downloading the last good feed.
Cost cap
Every feed has a monthly AI budget. We get a warning at 80%, and when it runs out, the run stops: no surprises on the bill.
Approval
We start with a pilot: you approve every value in the first sample. Anything still uncertain later never goes out on its own.
Reversible
Every feed version is kept, and we can roll back to any of them in one step. We never touch your original feed.
One dataset, every channel
The Google, Meta, ChatGPT, Pinterest and TikTok feeds are all built from the same verified data, each in its own format.
If any of these sound familiar…
Online stores
More attributes filled, more accurate search matches and a place in AI answers, without rebuilding your store.
Agencies
All your clients in one place, with user roles, a client portal and your own branding (white label).
Marketing leads
An AI readiness score before and after, control-group measurement and a monthly report on the real impact.
Reliable data, not just “AI-generated text”.
Plenty of tools can fill a missing field. We make sure every value comes from the store's own pages, and we show you where.
Evidence for every value
Every value comes with a verbatim quote from the product page, and code checks that the quote is really there. If it isn't in the source, it doesn't go out.
Prohibited claims filtered
Health claims, guarantees, overpromises: filtered per category, with an optional second AI review. That lowers the risk of Google disapproving your products.
Client portal
The portal shows every change per product with its source quote. If you like, you approve the uncertain values there, and you can download everything as CSV or JSON.
Control-group measurement
Part of the catalog stays unchanged for a while, so the real impact is visible. You get a monthly report on coverage and feed health.
Multiple countries
Feeds in Hungarian, Czech, Slovak, Romanian, Polish, German and English from your localized pages. No machine translation: Google requires the feed and the landing page to be in the same language.
Rules and labels
Custom labels, title additions and category rules, without AI. Your manual decision always wins over a rule or the AI.
How is it different from what you use today?
| Rule-based feed tool | AI copywriting tool | Manual work | feedlove | |
|---|---|---|---|---|
| Fills missing attributes? | Only what's already in the feed | Yes, but it can make things up | Yes, slowly | Yes, from the product page |
| Where does the data come from? | Your existing feed | The AI's “knowledge” | Wherever a colleague copies it from | The store's own page, with a quote |
| Verification | Format only | A prompt: “don't make anything up” | Spot checks | Code checks the quote and the numbers |
| Prohibited claims | – | Usually none | Up to the colleague | Filtered per category |
| Updates | Automatic | Started by hand | Rarely | Daily, automatically |
| Impact measurement | – | – | – | Control group, monthly report |
You don't need to replace your current feed tool (e.g. Channable, DataFeedWatch): we add a supplemental feed next to it.
Works with any store that has a Merchant Center feed: WooCommerce, Shopify, Shoprenter, Unas and custom platforms.
Not another tool to learn. We do it for you.
feedlove is a managed service: we set it up, monitor it and fix it. You only approve, if you want to.
Merchant Center setup
We add the supplemental feed and check that Google fetches it correctly.
A dedicated contact
One person who knows your store, your rules and your categories, and whom you can reach directly.
Continuous monitoring
If a run is missed, the feed goes stale or a price differs from the product page, we get an alert and act on it.
Monthly review
A monthly report on coverage and feed health, with recommendations for the next step.
We start with a free audit.
The first step costs nothing: we show you what we can do with your own products. Then we quote a monthly fee tailored to your store.
1. Free audit
A quality check and an AI readiness score, with a real sample from your products. 1–2 days.
2. Pilot
Part of your catalog with approval: you see every value before it goes live.
3. Monthly fee
Based on the number of products, channels and countries, and on whether you want the client portal, approvals or white-label branding.
Frequently asked questions
Do you need to change our store or our existing feed?
No. We read your existing feed URL and your public product pages. You add a supplemental feed in Merchant Center, and we help you set it up. Your original feed stays untouched.
What happens if the AI gets something wrong?
A value only goes out if its quote is found on the product page and its numbers and names match the source. When the page contradicts itself (for example, two different sizes on the same page), the value waits for approval. You can override or hide anything, and every earlier feed version can be restored.
Does AI-written text have to be labeled?
Yes, where Google requires it, we label it. AI-written descriptions and titles go into the fields Google designates for them (structured_description, structured_title), with the required label, next to your original text.
Do you guarantee that ChatGPT or Google will recommend my products?
No, nobody can guarantee that: what gets recommended is decided by Google and the AI providers. What we make sure of is that your feed carries as much accurate, verified data about your products as possible, because that's what they work from. We measure the impact with a control group.
What is FeedloveBot, and do we need to allow it?
Our bot fetches your product pages politely and respects robots.txt. If robots.txt blocks it, two lines allow it. Details: the bot page.
Do we have to replace our current feed tool?
No. Your current feed tool (Channable, DataFeedWatch, CTX Feed or your platform's own feed) stays; we add a supplemental feed next to it. Merchant Center merges the two.
Do we need a developer on our side?
No. All we need is a feed URL, plus adding a supplemental feed in Merchant Center, which we help with. If robots.txt blocks our bot, a two-line allow rule is needed.
What results can we expect?
We measure instead of promising: part of your catalog stays unenriched for a while (a control group), so the monthly report shows the difference in your own store's numbers. The audit shows in advance how much coverage and the AI readiness score improve.
How long does it take to start?
The free audit takes one or two days: we show you on a sample of products what we can fill and how much the AI readiness score improves. Then comes a pilot with approval, followed by the live feed, updated daily.
What happens to the data if we cancel?
You can download every value we produced as CSV or JSON, together with the source quotes. The data is yours.
Let's look at your feed.
On a sample of your products, we show you what we can fill from your own product pages and how much your feed improves.
- ✓ Published to the feed1,834
- ⏸ Awaited approval235
- ✕ Filtered out (no evidence)158
- Basic quality check: titles, images, identifiers, categories
- AI readiness score now and after enrichment
- A real sample from your products, with source quotes
- No commitment