Get started with Sichta in seven short videos

Open Sichta from the Apps section of your admin and press Run scan. Sichta fetches your public storefront the way an AI shopping agent would and runs 26 deterministic checks.

• A readiness score from 0 to 100 with a Ready, Needs work or Critical badge.
• Four pillars: agent plumbing, structured data, catalog completeness, buyer-policy clarity.
• Answerability at a glance: how many of the 10 buyer questions your products can answer.

TipOn a development store, save your storefront password under Settings › Preview mode first. A password-protected store blocks every scan, because agents cannot see it either.

The Checks page lists all 26 checks with a Pass, Warn or Fail badge and a one-line detail of what Sichta saw. Every non-passing check has a How to fix section written for that check.

• Checks are grouped A to D, matching the four pillars on the dashboard.
• The detail line quotes what was found, for example which products fail.
• After you fix something and rescan, the check shows a Fixed and Verified date.

Catalog shows which products fail which checks. Missing image alt text is the most common issue, and Sichta can fix it for you from product titles.

• The confirmation says exactly what will be written to your store before it happens.
• One-click fixes are journaled and can be undone from Settings › Fix history within 30 days.
• Run a scan afterwards to see the check turn green.

TipOne-click fixes are available on Starter and above. On Free you still see the list of affected products.

Agent View renders any product page as text and data only: readable words, headings, JSON-LD structured data and identifiers. No design, no images, just what an agent gets.

• Left column: what shoppers see. Right column: what an agent can read.
• Counts of words, tables, images and images with empty alt text.
• Every gap links to the check that closes it.

llms.txt is a plain-text guide to your store that AI crawlers read first. Shopify already serves a default at your store root; Sichta generates a richer version from your live catalog and policies that you can preview, copy or download.

• Generate, then Regenerate any time your catalog changes. Starter and above keeps it fresh automatically.
• Paste the content into templates/llms.txt.liquid in your theme's code editor to publish it in place of Shopify's default.
• agents.md works the same way and describes entry points and rules of engagement.

Buyers ask AI agents the same ten things: price, stock, variants, shipping, returns and so on. Answerability measures whether each one can be answered from your machine-readable data alone.

• A store average out of 10 and a score for every sampled product.
• Per question: how many products can answer it, and which ones cannot.
• Each question links to the check that fixes it.

Answer Snapshots ask real AI engines a fixed set of questions about your store and keep the answers, so you can see whether you are mentioned and how that changes over time.

• Default questions are generated from your shop name and product type. Add your own on paid plans.
• Each answer is tagged: store mentioned or not, and changed since last snapshot.
• Snapshots are observations, not rankings. Use them for the trend.

TipSnapshots take a few minutes because every question goes to every engine in turn. Start one, do something else, come back.