Noibu

The Noibu connector gives merchants a turnkey way to ask "why" questions about their online store; traffic, behavior, conversion, performance, and errors, directly inside Claude, and act on the answers without leaving the conversation.

Once connected, Claude has live, read access to the merchant's Noibu data: sessions, page visits, customer journeys, conversion funnels, heatmaps and scrollmaps, session replays, JavaScript and network errors, Core Web Vitals, marketing attribution, and release events. Teams ask questions in plain English: "why did mobile checkout drop last week?", "which campaigns send bouncing traffic?" — and get answers grounded in their actual store data, not generic AI guesses.

Key features:

• Ecommerce-native data model — funnels, cart abandonment, checkout drop-off, PDP/PLP performance, AOV, and revenue per session are first-class concepts, so answers map to real business questions.
• Full-funnel analytics on demand — conversion rate, bounce rate, revenue, payment and delivery method mix, drop-off between any two funnel steps, segmented by device, country, channel, campaign, or custom attribute.
• Technical health diagnosis — priority errors ranked by impact, AI-assisted diagnoses, source-mapped stack traces, and Core Web Vitals (LCP, CLS, INP) per page, so teams go from symptom to root cause in one conversation.
• Rich inline visuals — interactive click/scroll heatmaps, conversion-funnel charts, and session-replay video rendered directly in the conversation.
• Scoped write actions — Claude can update issue workflow states (e.g., mark an error fixed or ignored) and log release events, only when explicitly asked. Everything else is read-only; no settings, billing, or configuration access.
• Privacy by design — PII is masked/removed at ingestion into Noibu, so no personal shopper data can be retrieved through the connector.