BetaFirst 10 companies · 2,000 free credits ›

One keyword in.
A spec sheet out.

AI for new-product development on Amazon. It finds a direction, breaks down the competition and writes the spec sheet, covering the days of research that slow every launch down. Every conclusion shows its evidence, and you decide whether to act on it. After launch, it also makes your listing images and watches the market every day.

Built by Jason Huang, former Tech Lead in Amazon retail search. He runs every demo and handles support personally.

talar.eporo.ai
Bamboo drawer organizerMarket assessment · Segment: Bamboo
Selection criteria passed13 / 18
Top-20 monthly revenue$1.24M
Largest brand share18.0%
Top problems in negative reviewsShare · sales-weighted
Doesn't fit the drawer31.0%
Bamboo cracks or splinters22.4%
Dividers slide around14.1%
Today's brief · 10-312 major changes

A competitor cut its price 15%. Its effective price is now below yours.

Spec sheet readyspec_sheets.xlsx
Market map·Keyword curation·18-criteria scoring·Top-20 competitor breakdown·Spec sheet and RFQ·AI-planned image sets, made image by image·Daily brief·Keyword rankings·Weekly ad review·

One tool,
from idea to launch

The four modules follow a product's lifecycle. A segment you pick in Market Map goes into Market Research in one click. Before launch, Creative Center makes the full set of listing images. After launch, Product Monitor watches your listing and your competitors every day.

01 · Market Map

See the market,
then pick a niche

Break a top-level category into three layers (market, category and segment) and see search volume, year-over-year share and seasonality.

  • Segments gaining share come first
  • Holiday segments flagged, so you don't stock up as the season ends
  • Like one? Open it as a research project in one click
About Market Map→
Home Storage › Drawer OrganizersYoY share · pts
Bamboo+1.8
Fabric+0.6
Acrylic−0.9
Holiday giftNot now
02 · Market Research

One keyword in,
a spec sheet out

Keyword curation, segment scoring and a breakdown of the top 20 competitors, ending in a spec sheet and RFQ written to your positioning.

  • Every keyword kept comes with a reason
  • Scored on 18 criteria, with a one-line verdict
  • Weight, carton specs and fees calculated from the spec
About Market Research→
Segment assessment · Bamboo18 criteria
72%13 of 18 product-selection criteria passedNeeds further evaluation
RFQ sheet38 × 26 × 12 cmCalculations checked
03 · Creative Center

Assets and images,
one workspace

Keep all your product assets in one place. Image generation uses your real assets as references, and the whole set of listing images is made in the same place.

  • Assets organized by product, no more hunting for files
  • Images automatically use your real product photos and selling points
  • An AI director plans the set; an AI designer makes each image
About Creative Center→
1 Main
2 Size
3 Benefits
4 Scene
5 Scene
6 Material
7 Versus
AI designerGenerating…
8 Install
Queued
9 Packing
04 · Product Monitor

Daily action guide

Watches your listings and your competitors, and builds ad performance reports with suggested adjustments. Whether to change anything, and how, is yours to confirm.

  • Quiet days get a one-line brief
  • Organic and sponsored ranks tracked together
  • Once you make a change, it's marked as done automatically
About Product Monitor→
Fri 10-312 major changes

A competitor cut its price 15%. Its effective price is now below yours.

Price gap+$1.20
Organic#7
Rating4.5
4 ad suggestions this weekBid changes wait for your confirmation

AI analyzes.
You decide.

Every number has a source

Reports export to Excel, and every column says where the number comes from and how it was calculated. Forward it to a colleague, or send it straight to a factory.

Every change needs your approval

Talar makes suggestions. Any change to prices, listings or ads needs your approval; without it, your store is not touched.

Also: interface and reports in English and Chinese · Admin, member and read-only roles · Cost shown for every project

Real case · Cross-border pet-supplies seller

Talar was first built together with this seller, following their existing product-selection process.

Market research, competitor breakdown, writing the spec and finding a factory, moved step by step into one tool.
< 10 minMachine run time per research run
Measured
2+ daysTime saved per research run on
competitor research and data preparation · Customer feedback

Who builds TalarTalar is designed and built by Jason Huang, founder of Eporo. He was a Tech Lead at Amazon, working on the backends for retail search and product detail pages. He runs the demo and the support that follows himself.

Full profile→

FAQ

How does the beta work?

The beta is open to the first 10 companies only. Leave your WeChat ID or email and we'll contact you to book a 30-minute demo. After the demo, you buy credits and connect your store to get started; beta accounts get 2,000 free credits.

How is it priced?

By usage: each task uses credits. After the demo you buy credits and connect your store to get started, and each company in the beta receives 2,000 free credits. We'll give you the per-credit price in the demo.

Which accounts do I need?

Authorization for your Amazon store, plus your Amazon Ads account if you want ad reports. Only Market Map needs Amazon Brand Analytics access, which requires a brand enrolled in Amazon Brand Registry; Market Research, Creative Center and Product Monitor don't. We'll go through each item with you in the demo.

Which marketplaces are supported?

The US marketplace for now. We'll confirm other marketplaces with you in the demo.

Is my data safe?

Each company's data is stored separately, and other customers can't see it. Eporo does not train models on your data. If you stop using Talar, you can ask us to delete your data at any time.

How does this relate to the data tool I use now?

Talar includes the market data it needs, so you don't need to renew a separate data tool for it. A data tool gives you numbers; Talar organizes them, makes a judgment and gives recommendations. What it saves is the days it takes to turn numbers into conclusions.

What if the AI gets it wrong?

Every conclusion comes with its evidence, so you can check it point by point. Key decisions, such as whether to go ahead, how to differentiate the product and whether to change your store, are yours to make.

Can I use it from an AI chat?

Yes. Connect Talar to the AI assistant you already use, and you can open projects, read reports and ask about the conclusions there.

What happens in the demo?

Bring whatever fits the module you care about most: a category link for Market Map, a keyword and a few reference products for Market Research, a few product photos and selling points for Creative Center, or a few ASINs for Product Monitor. We run it live. It takes 30 minutes, it's free, and the founder runs it himself.