Introspect Operations

Observe system state, inspect records, forecast decay, and monitor evolution.

introspect.stats

Retrieve system-wide statistics including record counts, average fidelity, background decay status, and optional evolution metrics.

Endpoint: GET /v1/introspect/stats

Response

json

{
  "total_records": 342,
  "records_by_store": {
    "episodic": 150,
    "semantic": 120,
    "procedural": 35,
    "emotional": 25,
    "working": 12
  },
  "total_associations": 890,
  "avg_fidelity": 0.742,
  "avg_fidelity_by_store": {
    "episodic": 0.68,
    "semantic": 0.89,
    "procedural": 0.91,
    "emotional": 0.82,
    "working": 0.45
  },
  "oldest_record": "2026-01-15T08:30:00Z",
  "newest_record": "2026-03-19T09:45:00Z",
  "total_recall_count": 1567,
  "raw_journal_records": 1280,
  "raw_journal_pending_dream": 45,
  "dream_episodic_summaries": 60,
  "dream_semantic_nodes": 25,
  "last_dream_tick_at": "2026-03-19T06:00:00Z",
  "background_dream_enabled": true,
  "vector_index_records": 320,
  "vector_search_backend": "brute_force",
  "evolution_metrics": {
    "parameter_adjustments": [],
    "detected_patterns": [],
    "schema_adaptations": []
  },
  "background_decay_enabled": true
}

Field	Type	Description
`total_records`	u32	Total records across all stores
`records_by_store`	Map	Record count per store
`total_associations`	u32	Total association edges
`avg_fidelity`	f64	Average fidelity score across all records
`avg_fidelity_by_store`	Map	Average fidelity per store
`oldest_record`	DateTime	Timestamp of the oldest record
`newest_record`	DateTime	Timestamp of the newest record
`total_recall_count`	u64	Total recall operations performed
`raw_journal_records`	u32	Total raw journal entries
`raw_journal_pending_dream`	u32	Raw journal entries not yet processed by dream tick
`dream_episodic_summaries`	u32	Episodic summaries created from dream processing
`dream_semantic_nodes`	u32	Semantic nodes promoted from dream processing
`last_dream_tick_at`	DateTime or null	Timestamp of the last dream tick
`background_dream_enabled`	bool	Whether background dream processing is running
`vector_index_records`	u32	Number of records indexed in the redb-backed vector store
`vector_search_backend`	string	Active vector search backend identifier (currently `brute_force` — deterministic redb cosine scan)
`evolution_metrics`	EvolutionMetrics or null	Adaptive tuning metrics when available
`background_decay_enabled`	bool	Whether background decay is running

introspect.record

Inspect a specific memory record with full detail including fidelity state, associations, and access history.

Endpoint: GET /v1/introspect/record/:record_id

Request Parameters

Parameter	Type	Description
`record_id`	UUID (path)	The record to inspect

The CMP request model also defines include_history, include_associations, and include_versions. The current HTTP transport always returns the standard record shape and does not expose those toggles as query parameters.

Response

Returns the full MemoryRecord as JSON:

json

{
  "id": "01916e3a-5678-7000-8000-000000000001",
  "store": "episodic",
  "created_at": "2026-03-15T14:30:00Z",
  "updated_at": "2026-03-15T14:30:00Z",
  "last_accessed_at": "2026-03-18T10:00:00Z",
  "access_count": 5,
  "content": {
    "blocks": [
      {
        "modality": "text",
        "format": "text/plain",
        "data": [72, 101, 108, 108, 111],
        "embedding": [0.1, 0.2, 0.3]
      }
    ],
    "summary": "User greeting"
  },
  "fidelity": {
    "score": 0.847,
    "noise_level": 0.12,
    "decay_rate": 0.3,
    "emotional_anchor": 1.0,
    "reinforcement_count": 5,
    "stability": 8.42,
    "last_decay_tick": "2026-03-19T09:00:00Z"
  },
  "emotion": {
    "joy": 0.0,
    "trust": 0.8,
    "fear": 0.0,
    "surprise": 0.0,
    "sadness": 0.0,
    "disgust": 0.0,
    "anger": 0.0,
    "anticipation": 0.0,
    "intensity": 0.6,
    "valence": 0.7
  },
  "associations": [
    {
      "target_id": "01916e3a-9999-7000-8000-000000000002",
      "association_type": "semantic",
      "weight": 0.85,
      "created_at": "2026-03-15T14:30:00Z",
      "last_co_activation": "2026-03-18T10:00:00Z"
    }
  ],
  "metadata": {},
  "version": 1
}

introspect.decay_forecast

Predict the future fidelity of specific records at a given point in time.

Endpoint: POST /v1/introspect/decay-forecast

Request

json

{
  "record_ids": [
    "01916e3a-5678-7000-8000-000000000001",
    "01916e3a-5678-7000-8000-000000000002"
  ],
  "forecast_at": "2026-04-19T00:00:00Z"
}

Field	Type	Required	Description
`record_ids`	Uuid[]	Yes	Records to forecast
`forecast_at`	DateTime	Yes	Target date for the forecast

Response

json

{
  "forecasts": [
    {
      "record_id": "01916e3a-5678-7000-8000-000000000001",
      "current_fidelity": 0.847,
      "forecasted_fidelity": 0.312,
      "estimated_threshold_date": "2026-06-15T00:00:00Z"
    },
    {
      "record_id": "01916e3a-5678-7000-8000-000000000002",
      "current_fidelity": 0.923,
      "forecasted_fidelity": 0.671,
      "estimated_threshold_date": null
    }
  ]
}

introspect.evolution

View the evolution engine's adaptive parameter adjustments, detected usage patterns, and schema adaptations.

Endpoint: GET /v1/introspect/evolution

Response

json

{
  "parameter_adjustments": [
    {
      "store": "episodic",
      "parameter": "decay_rate",
      "original_value": 0.3,
      "current_value": 0.25,
      "reason": "High recall frequency detected"
    }
  ],
  "detected_patterns": [
    "frequent_semantic_recall",
    "low_working_memory_usage"
  ],
  "schema_adaptations": []
}

Next Steps

API Endpoints

Invoke these operations over HTTP

CLI Reference

Invoke these operations from the CLI