Amami MCP docs

Natural language queries

Natural language queries are the main reason to connect Amami to an AI assistant. You describe the analytics question; the assistant chooses tools, fills parameters, reads results, and explains the answer.

Query types

Basic traffic:

How many visitors did the site get yesterday?

Content performance:

Which pages performed best in the last 30 days, and what do they have in common?

Source quality:

Compare GitHub, Google, X, newsletter, and Product Hunt traffic. Which source is most likely to reach signup?

Device and country analysis:

Show me whether mobile visitors behave differently from desktop visitors this week.

Event and conversion analysis:

Which conversion events changed the most this month?

Time comparison:

Compare this week with last week. What changed, what caused it, and what should I do next?

Realtime:

How many people are online right now, where did they come from, and which page are they viewing?

Write better prompts

Include:

  • Website or domain.
  • Date range.
  • Metric or behavior.
  • Comparison period or segment.
  • Output format.
  • Decision you are trying to make.

Weak:

How is my site doing?

Better:

Analyze docs.amami.dev for the last 14 days. Compare top pages, referrers, and signup-related events with the previous 14 days. Give me three actions ranked by expected impact.

Ask for evidence

The assistant should not give generic advice. Ask it to include:

  • Tool results used.
  • Date range.
  • Metric values.
  • Segment names.
  • Confidence level.
  • Follow-up measurement plan.

Prompt:

Answer with evidence from Amami tool results. If the data is missing, tell me which event or page I need to track.

Useful query templates

Daily check:

Summarize yesterday's visitors, pageviews, top pages, top referrers, active campaigns, and unusual changes.

Content review:

Find the five pages with the highest growth this month. Explain why they grew and what related page I should publish next.

Funnel review:

Review the signup funnel. Find the largest drop-off, compare by device, and recommend one product or copy change.

Instrumentation review:

Look at my pages and events. Which important conversion events are missing?