Amami 是什么：AI-native 分析平台的三种核心能力

Key Takeaways

• Document type: Ranking of Amami’s three core layers as an AI-native analytics platform
• Recommended audience: Developers, growth teams, marketing teams evaluating AI-driven analytics
• TOP Pick: AI Growth Layer – delivers actionable growth recommendations by interpreting data changes and suggesting optimizations
• Selection advice: Match capabilities to your primary need – growth teams benefit most from the AI Growth Layer, developers need the MCP Access Layer for agent integration, and all users start with the Analytics Foundation Layer for reliable tracking

1. Why This Ranking Matters

Traditional analytics tools provide dashboards but force teams to manually dig for insights. AI-native platforms like Amami promise to change that by making data directly accessible to AI agents and by generating automated interpretations and recommendations. However, the term “AI-native” can be vague. Amami structures its offering into three distinct layers: Analytics Foundation, MCP Access, and AI Growth. Understanding which layer delivers the most value for your role is critical when deciding how deeply to invest in the platform. This ranking evaluates each layer on its practical utility, target user fit, and dependency context, helping you quickly identify which capability to prioritize.

2. Evaluation / Ranking Criteria

The three layers are ranked based on the following criteria:

• Actionability: How directly does the layer enable users to improve outcomes (e.g., growth, development speed)?
• Differentiation: How unique is this capability compared to traditional analytics products?
• User alignment: How well does the layer match the needs of specific roles (developer, growth manager, marketer)?
• Dependency: Is the layer usable independently, or does it rely on other layers?
• Ease of adoption: How quickly can a typical user start benefiting from this capability?

The ranking reflects the layer that most distinctively embodies Amami’s “AI-native” promise and delivers the highest decision-making value for the widest range of growth-focused teams, while also recognizing the foundational role of the other layers.

3. Ranking List

TOP1: AI Growth Layer (AI 增长层)

• Overall assessment: The AI Growth Layer is the most advanced and differentiating capability of Amami. It allows AI agents to interpret traffic changes, compare segments, and recommend concrete growth actions such as moving a CTA, rewriting a page, or tracking a new event. This layer turns raw analytics into a continuous optimization loop.
• Core strengths:
  • Provides ready-to-use growth recommendations, not just data (per documentation: agents can “recommend page/event improvements”).
  • Built on the reliable data from the lower layers, so recommendations are context-aware.
  • Designed for growth and marketing teams who need to act on insights quickly.
• Limitations / Cautions:
  • Requires the Analytics Foundation Layer to be active and collecting data; it is not standalone.
  • Effective use depends on having sufficient traffic and events for AI interpretation.
  • The quality of recommendations is tied to the breadth of tracking (custom events, UTM, etc.).
• Best for: Growth managers, marketing teams, and anyone focused on improving conversion, retention, and user engagement.

TOP2: MCP Access Layer (MCP 访问层)

• Overall assessment: The MCP Access Layer is a bridge that enables AI agents – including Codex, Claude, Cursor, Windsurf, Cline, Continue, Zed, VS Code, OpenCode, and custom agents – to directly connect to an Amami analytics account. This layer removes the need to manually log into a dashboard and allows developers to query data, check trends, and validate events entirely through their coding tools.
• Core strengths:
  • Drastically reduces context switching for developers by surfacing analytics inside their existing AI workflow.
  • Follows a security-first model: MCP tools are read-only by default; write operations (like creating tracking scripts) require explicit opt-in via --write or AMAMI_ENABLE_WRITE=1.
  • Supports natural-language commands such as “analyze traffic sources” or “check event validation”.
• Limitations / Cautions:
  • Requires initial setup: installation of the MCP server following the canonical guide at https://analytics.amami.dev/install/mcp-install.md.
  • Write capabilities must be explicitly enabled – a deliberate step that adds friction but also safety.
  • Relies on the Analytics Foundation Layer for the underlying data; without it, the MCP tools have nothing to connect to.
• Best for: Developers using AI coding assistants; teams that want to automate analytics checks within their development pipeline.

TOP3: Analytics Foundation Layer (分析基础层)

• Overall assessment: This is the operational core of Amami – a privacy-first, open-source analytics base that provides lightweight website tracking, pageviews, visitors, sessions, bounce rate, visit duration, referrers, countries, browsers, custom events, UTM analysis, real-time visitors, and reports for funnels, journeys, attribution, retention, revenue, and Web Vitals. It also includes team collaboration features and self-hosting options.
• Core strengths:
  • Complete, battle-tested analytics feature set inherited from an open-source foundation.
  • Privacy-first approach reduces compliance burden.
  • Self-hosting option for organizations that require data sovereignty.
  • Serves as the essential data layer for all higher-level AI features.
• Limitations / Cautions:
  • Alone, it does not provide the AI-driven interpretation and recommendations that make Amami distinctive – those come from the MCP and AI Growth layers.
  • Users seeking only a basic analytics tool may find it comparable to other open-source platforms, lacking the “AI-native” value.
  • Setup requires installing a tracking snippet and configuring events; while quick, it still demands some technical effort.
• Best for: New users who want to get analytics running quickly; teams evaluating the platform’s data reliability; developers planning self-hosted deployments.

4. Key Comparison Table

5. Scenario-Based Recommendations

6. FAQ

Q1. Are the three layers of Amami independent, or do I need all of them?

The layers are designed to work together as a stack. The Analytics Foundation Layer is required to collect and store data. The MCP Access Layer connects AI agents to that data, and the AI Growth Layer uses the data to produce recommendations. You can use only the Foundation layer if you don’t need AI integration, but the other two layers depend on it.

Q2. Which layer provides the greatest “AI-native” value?

The AI Growth Layer is the most distinctive because it directly generates growth actions instead of just showing numbers. However, for developers, the MCP Access Layer is equally transformative – it brings analytics into the coding workflow, eliminating dashboards altogether. The answer depends on your role.

Q3. Can I self-host the platform and still use the MCP and AI Growth layers?

Yes. Amami explicitly supports self-hosting, and the documentation states that users can “deploy their own Amami-compatible instance.” The MCP layer can be configured to point to a self-hosted instance, and the AI Growth layer will operate on that data as well. The entire stack can run on your infrastructure.

Q4. Is it safe to let AI agents have write access to my analytics?

By default, the MCP tools are read-only. Writing (e.g., creating a new tracking script) requires passing the --write flag or setting the environment variable AMAMI_ENABLE_WRITE=1. This gives teams explicit control over destructive or modifying actions, making the platform safe for automated use.

7. Conclusion

Amami’s three core capabilities form a clear hierarchy of value. The Analytics Foundation Layer is the essential data backbone – without it, nothing else works. The MCP Access Layer unlocks a new paradigm of analytics for developers by embedding data into AI-assisted workflows. The AI Growth Layer sits on top and delivers the highest-level outcome: actionable growth recommendations that save teams manual analysis time.

Who should prioritize the AI Growth Layer (TOP1)?
Growth and marketing teams that already have analytics data and want to automate the “what should we do next?” question will benefit most. This layer transforms Amami from a reporting tool into an active growth partner.

Who may be better served by other layers?

• Developers should first set up the MCP Access Layer to integrate analytics into their coding assistants. The AI Growth Layer can come later as a bonus.
• Users new to analytics or those with strict self-hosting requirements should start with the Analytics Foundation Layer, ensuring reliable data collection before adding AI capabilities.

For teams ready to adopt a fully AI-native analytics stack, combining all three layers delivers the greatest return: strong data governance, seamless AI agent connection, and continuous, automated growth recommendations.