get_memory_health

Get a consolidated health report across all 5 layers for a project — one call instead of five.

Overview

get_memory_health aggregates diagnostics from every layer of the Wake Intelligence brain into a single response. It replaces the need to call get_memory_stats, get_causality_stats, get_propagation_stats, get_learning_stats, and get_causal_graph separately when you just want a project health check.

Added: v3.5.0

Layer: All 5 layers (consolidated diagnostic)

Purpose: Single-call health check — context distribution, causal coverage, predictions, learning status

---

Parameters

Parameter	Type	Required	Description
project	string	Yes	Project identifier

---

Returns

A formatted health report covering all layers:

**Wake Intelligence Health Report — api-service**

🧠 **Layer 1: Causality**
  - Contexts with causal metadata: 34 / 47 (72%)
  - Causal chains: 8
  - Average chain depth: 3.2

📊 **Layer 2: Memory Tiers**
  - ACTIVE: 3
  - RECENT: 12
  - ARCHIVED: 28
  - EXPIRED: 4

🔮 **Layer 3: Predictions**
  - Contexts with prediction scores: 31
  - Average score: 0.64
  - High-value (score > 0.7): 11

🎓 **Layer 4: Meta-Learning**
  - Outcomes recorded: 43
  - Learning active: Yes (requires ≥20)
  - Temporal weight: 0.52 | Causal: 0.28 | Frequency: 0.20

🎭 **Layer 5: Personality**
  - Available modes: historian, prophet, archaeologist, minimalist, auditor
  - Last used mode: historian

---

Example

get_memory_health({ project: "wake-intelligence" })

---

Use Cases

Cold Boot Health Check

Start a new session by running get_memory_health to immediately understand the state of a project — what's active, what's expired, whether predictions are running, and whether learning has kicked in.

Debugging Search Quality

If search_context is returning poor results, the health report shows whether Vectorize embeddings exist (via prediction coverage) and whether contexts are in the right memory tiers.

Monitoring After admin_reindex_all

Run get_memory_health before and after reindexing to confirm Vectorize coverage improved.

---

vs. Individual Stats Tools

	get_memory_health	Individual tools
Call count	1	4-5
Detail level	Summary per layer	Full detail per layer
Best for	Health checks, cold boot	Deep-diving a specific layer

---

See Also

• get_memory_stats — Full Layer 2 memory tier breakdown
• get_causality_stats — Full Layer 1 causality analytics
• get_propagation_stats — Full Layer 3 prediction analytics
• get_learning_stats — Full Layer 4 weight tuning detail
• admin_reindex_all — Fix empty Vectorize index