The connector list — and the two names not on it#

Live Artifacts ships with ten native connectors. Two of the most common WooCommerce data sources aren’t among them, and that’s the entire story.

When Anthropic launched Live Artifacts inside Claude Desktop on April 20, 2026, the connector library went live with PostHog, Mixpanel, Datadog, Airtable, Apify, Gmail, Shopify, Slack, Zapier and a handful of others. Two names that were not on the list, and remain not on the list as of this writing: Google Analytics 4 and Meta Ads.

The Anthropic public roadmap has not committed a date for either. The May 2026 practitioner audit by Annika Helendi was the first widely-circulated piece to call this out directly, and the absence has been confirmed in Anthropic’s own support documentation since.

Two of the most queried analytics tools on the planet are not natively reachable from the most-talked-about AI dashboard interface of 2026. That is not an oversight — that is an architectural fact a WooCommerce operator has to plan around.

For Shopify operators this is a non-event — the Shopify connector ships native, and dashboards build straight from it. For WooCommerce operators it is the opposite. The platform that 43.5% of the web runs on doesn’t have a native connector either, and the two analytics destinations a WooCommerce store usually leans on for ad performance are missing too.

The architectural decision the missing connectors force#

If Claude Desktop can’t reach GA4 or Meta Ads, there are only four places ad and conversion data can live — and three of them are dead ends.

List the options a WooCommerce store actually has, and the choice collapses fast.

• Keep it in GA4. No native connector. Claude Desktop can’t read it. Dead end.
• Keep it in Meta’s Ads Manager. No native connector. Same dead end.
• Keep it in WooCommerce admin. The Shopify connector exists, but there’s no WooCommerce equivalent. Dead end for native use.
• Keep it in BigQuery. Google’s remote BigQuery MCP server reached general availability in early May 2026, authenticated via OAuth 2.0 with no API keys required. This is the only option that works.

By elimination, BigQuery is the data home — not because it is the best in some abstract sense, but because it is the only one Claude Desktop can natively reach. The “where does our data live?” question has been settled by an absence on someone else’s connector roadmap.

You may be interested in: How WordPress Events Reach BigQuery in Seconds

What BigQuery actually has to contain for this to work#

Getting data into BigQuery is the easy half. Getting the right data into BigQuery is where most warehouses fail the Live Artifact test.

The temptation, on hearing “BigQuery is now the data home”, is to assume the existing WooCommerce-to-BigQuery ETL plugin is enough. It is not. The standard WooCommerce REST API exports six entity types: orders, products, customers, coupons, refunds and shipping. Every one of those is a snapshot of a completed transaction. None of them is the behavioural event stream Live Artifacts will be asked to reason over.

For a Live Artifact to answer the kinds of questions a 2026 operator wants — “which campaign drove revenue this hour?”, “which products are getting added to cart but not bought today?”, “what’s our ROAS by campaign in the last 24 hours?” — the BigQuery dataset has to carry:

• Event-level rows, not order-level rows. Every page view, every add to cart, every begin checkout — with timestamps in microseconds.
• UTM parameters stamped on the same row as the event. Attribution can’t be a separate join job at query time; it has to be already there.
• Stable user IDs across sessions. A returning customer needs to be the same row identifier today as last month.
• Conversion events joined to ad spend at session granularity. Hourly ROAS is impossible without it.
• Indefinite retention. GA4’s free tier silently drops events older than 60 days from BigQuery export — making last-quarter queries impossible.

A typical WooCommerce ETL warehouse contains six tables and zero behavioural events. A warehouse that can answer Live Artifact questions contains the full enhanced ecommerce event spectrum, with attribution stamped at capture, and indefinite retention.

ETL warehouse vs server-side warehouse — side by side#

The same Live Artifact prompt produces “no data” against one warehouse and a full dashboard against the other. The difference is the schema.

Question a Live Artifact will be asked	Default ETL warehouse	Server-side event warehouse
Which campaign drove revenue in the last hour?	No — no campaign on order rows	Yes — utm_campaign stamped on purchase event
Which products are getting added to cart but not bought today?	No — no add_to_cart events	Yes — every add_to_cart event captured
What’s our ROAS per campaign this morning?	No — no spend join, no campaign attribution	Yes — conversion events joined to ad spend
Returning vs new customer split this week?	Partial — customers table exists, no session continuity	Yes — stable user_id across sessions
How many orders did Mary place last month?	Yes — orders table is fine for this	Yes — same
Top product velocity this hour vs same hour last week?	No — order grain is daily at best	Yes — microsecond event timestamps

Five of the six common questions return nothing useful from a default ETL warehouse, and the one that does work — counting completed orders — is the question store owners didn’t actually need an AI dashboard to answer. Looker Studio has done that for a decade.

You may be interested in: The GA4 BigQuery Event Limit — What Happens When Your WooCommerce Store Hits 1 Million

How to get there without rebuilding your stack#

The schema gap is solvable without re-platforming. Server-side event capture writes the missing rows directly into BigQuery alongside the ETL exports.

The work isn’t to abandon WooCommerce or rip out the existing ETL pipeline. The ETL exports of orders, products and customers are fine where they are. What’s missing is a parallel stream that writes the behavioural events — page views, add to cart, checkout steps, purchase events — into BigQuery the moment they happen, with attribution and user identity already stamped at capture.

Transmute Engine™ is a first-party Node.js server that runs on your subdomain (for example, data.yourstore.com). The inPIPE WordPress plugin captures the WooCommerce hooks and sends events via API to your Transmute Engine server, which streams the full enhanced-ecommerce spectrum directly to BigQuery — with utm_source, utm_campaign, user_id and session_id in the same row as the event.

The result is that the same Claude Desktop prompt that returned “no data” against an ETL warehouse now returns a working dashboard against a Transmute Engine warehouse. The connector library didn’t change. The data did.

Key Takeaways#

Five facts to leave with — the connector gap, the forced choice, and the schema work that makes Live Artifacts actually answer your questions.

• No native connector: Claude Desktop Live Artifacts has no GA4 or Meta Ads connector and no committed roadmap date.
• BigQuery by elimination: Of the four places WooCommerce ad data can live, only BigQuery is natively reachable by Claude Desktop.
• ETL is not enough: The default WooCommerce-to-BigQuery ETL syncs six entity types — orders, products, customers, coupons, refunds, shipping — and zero behavioural events.
• Schema is the moat: Five of six common Live Artifact questions need event-level rows with attribution stamped at capture.
• Server-side fixes it: A first-party event capture server writes the missing schema into BigQuery alongside the ETL data, without re-platforming the store.

Frequently Asked Questions#

Quick answers to the five questions store owners ask first when they hear Live Artifacts can’t natively reach GA4 or Meta Ads.

References#

Primary sources behind the claims above — platform documentation, official schemas and the May 2026 practitioner review that surfaced the connector gap.

• Anthropic. (2026). Live Artifacts: Connectors and integrations. support.claude.com
• Helendi, A. (2026, May). A practitioner’s audit of Claude Desktop Live Artifacts connectors. Substack.
• Google Cloud. (2026). BigQuery MCP server documentation. cloud.google.com
• WooCommerce. (2025). REST API reference. woocommerce.github.io
• W3Techs. (2024). Usage statistics of content management systems. w3techs.com
• Google. (2025). GA4 BigQuery export schema. developers.google.com

Server-side event capture into BigQuery is the foundation every 2026 Live Artifact will sit on top of. See how Transmute Engine writes the full enhanced-ecommerce spectrum direct from your WooCommerce store at seresa.io.