AI-Powered Insights
Stop staring at charts. Ask your data.
Weekly summaries in plain English. Anomaly alerts when something changes. An MCP server so AI agents like Claude can query your analytics directly.
Traffic up 14% — organic search driving growth, led by your GA4 events guide.
Pricing page dropped 3 positions — now at position 11 for “dashboard pricing”.
Product Hunt referral spike — +800 sessions Wednesday, back to baseline.
Weekly AI summaries
Coming in Phase 2Every Monday morning, Nook sends you an email with three bullet points: here is what happened on your site this week. Written in plain English, grounded in your actual data.
Nook Weekly Digest
Your analytics summary for Mar 10 – Mar 16
Traffic is up 14% week-over-week — driven mostly by organic search. Your post “How to Set Up GA4 Events” gained 2,400 new sessions from Google.
Your pricing page lost 3 positions for “analytics dashboard pricing” (now position 11). Impressions dropped 40%. Worth investigating.
Referral traffic from Product Hunt spiked on Wednesday (+800 sessions) and has since returned to baseline. Likely a one-time feature or mention.
Example summary — your real digest reflects your actual site data.
Anomaly detection
Coming in Phase 2Nook watches your metrics continuously. When something moves outside normal ranges — a traffic drop, a ranking change, a sudden spike — you get an alert with context, not just a number.
Traffic drop
Sessions fell 32% on Tuesday compared to the same day last week. The decline is concentrated on organic search traffic to your /blog/ pages.
Ranking change
Your homepage moved from position 3 to position 8 for "simple analytics dashboard". Impressions dropped 60% for this keyword.
Traffic spike
Referral traffic increased 5x on Thursday. The source is a link from a popular newsletter. The spike is concentrated on your /features page.
CTR improvement
Click-through rate for "ga4 alternative" jumped from 4.2% to 9.8% after your title tag change last week. Clicks are up 134%.
Example alerts — real anomaly detection uses your historical baselines.
MCP server
Available nowNook exposes your analytics through an MCP (Model Context Protocol) server. This means AI assistants like Claude can read your data directly and answer questions about your traffic, search performance, and content.
What you can ask
What were my top 10 pages by sessions last week?
Which search queries have the best CTR improvement this month?
Show me pages that lost more than 20% traffic compared to last month.
What are my striking distance keywords for the /blog/ section?
How it works
- 01
Generate an API key in your Nook settings.
- 02
Add the Nook MCP server to your AI assistant's configuration.
- 03
Start asking questions about your analytics in natural language.
Coming soon: Ask your data
Planned — BusinessType a question in the Nook dashboard, get an answer. Conversational analytics powered by AI, grounded in your real data. No query language to learn, no reports to configure.
Why did my traffic drop on Tuesday?
Your organic search traffic dropped 28% on Tuesday. The decline is concentrated on 3 blog posts that lost rankings: “GA4 Setup Guide” dropped from position 4 to 9, “Analytics Dashboard Review” from 6 to 14, and “Best SEO Tools” from 3 to 7. This looks like it could be related to a Google algorithm update — other ranking shifts started the same day.
Planned feature — available on the Business tier when it launches.
Feature availability by tier
| Feature | Starter | Growth | Business |
|---|---|---|---|
| MCP serverAvailable now | — | ||
| Weekly AI summariesComing in Phase 2 | |||
| Anomaly detectionComing in Phase 2 | — | ||
| Ask your dataPlanned | — | — |
Frequently asked questions
Nook uses a combination of statistical analysis for anomaly detection and large language models for generating plain-English summaries. The specific models may evolve as the technology improves, but the goal stays the same: turn numbers into understanding.
The summaries are grounded in your actual analytics data — they describe real numbers, not predictions. The AI layer translates metrics into readable language. If your sessions dropped 18%, the summary will say that and explain which sources or pages drove the change.
MCP (Model Context Protocol) is an open standard that lets AI assistants like Claude connect to external data sources. Nook's MCP server gives AI agents read-only access to your analytics, so you can ask questions like 'What were my top pages last week?' in a conversation.
Aggregated analytics data is processed to generate summaries. We never send raw user-level data to any AI provider. The data used for summaries is the same aggregated metrics you see in your dashboard — sessions, pageviews, queries, and trends.
The MCP server is available now on Growth and Business plans. Weekly AI summaries and anomaly detection are coming in Phase 2. Conversational analytics ('Ask your data') is planned for the Business tier. We will update this page as features launch.