Metrifyr
Memory System

Memory System

Persistent AI context per API key for improved responses

Memory System

The Metrifyr memory system stores context that persists across AI conversations, tied to each API key.

What is Memory?

Memory allows the MCP server to remember:

  • Your preferred GA4 properties
  • Common date ranges you use
  • Previous diagnostic findings
  • Custom preferences
  • Mistakes to avoid (learned from errors)

Each API key has its own memory, enabling different configurations for different projects.

Memory also learns automatically from your usage patterns, creating personalized context that improves over time.

How It Works

  1. MCP Server authenticates with your API key
  2. Metrifyr retrieves memory for that specific API key
  3. Context is injected into AI conversations
  4. AI provides more relevant responses

Benefits

Without MemoryWith Memory
Re-specify property IDsRemembered automatically
Generic responsesPersonalized to your data
No historical contextBuilds on past conversations
Same settings for all projectsPer-project configuration
Repeat same mistakesLearns from errors

Memory Architecture

Memory is tied to API keys, not users. This means:

  • Production API key can have production GA4 properties
  • Development API key can have test properties
  • Different clients can have different preferences

Two-Level Toggle

  1. Global toggle (header) - Master switch for all memory
  2. Per-key toggle - Enable/disable memory for each API key

Memory is active only when both are enabled.

Enabling Memory

  1. Create an API key in Dashboard
  2. Enable memory globally using the header switch
  3. Enable memory for specific API keys in Dashboard > Memory

Configuration Guide | AI Learning

On this page

Memory System | Metrifyr Docs | Metrifyr