Module models 

Core data models.

Modules

has_datapoint

The HasDataPoint trait used by the provenance-stamping algorithm in cognee-core::provenance.

memory

Memory entry types for typed remember() dispatch. Discriminated-union memory entries for remember() typed dispatch.

permission

Permission helpers.

temporal_event

Temporal event and timestamp types for the cognify pipeline.

Structs

BackendConfig

Configuration for dynamically selecting/creating a backend.

BackendOverrides

Per-call backend configuration overrides.

CognifyInterval

A time range stored as a graph node of type “Interval”. Mirrors Python: cognee.modules.engine.models.Interval Field names time_from / time_to match Python exactly.

CognifyTimestamp

A point in time extracted from text during temporal cognify. Mirrors Python: cognee.modules.engine.models.Timestamp time_at stores milliseconds since Unix epoch (UTC) — same unit as Python.

Data

Represents a piece of data in the system, such as a file or a text. Fields match the Python cognee data table schema for cross-SDK compatibility.

DataPoint

Base model for all storage-layer entities.

Dataset

Document

A classified document derived from a Data item.

DocumentChunk

A chunk of text extracted from a document during the cognify pipeline.

EdgeMetadata

Edge metadata for relationships between DataPoints.

EdgeType

Storage-layer edge type model.

Embedding

Embedding vector for a data point field.

Entity

Storage-layer entity model.

EntityType

Storage-layer entity type model.

EventAttribute

An entity related to an event, extracted during temporal entity enrichment.

FeedbackEntry

Feedback attached to an existing QA entry.

QAEntry

A Q&A turn stored in the session cache.

RawExtractedTimestamp

LLM output schema for a timestamp. Mirrors Python task model Timestamp. All fields except year default to 1/0; the extractor computes time_at.

TemporalEvent

An event extracted from text, optionally anchored to a point or range in time. Mirrors Python: cognee.modules.engine.models.Event

TraceEntry

One step of an agent trace.

Triplet

A triplet representing a semantic relationship between two entities.

User

A registered user. Corresponds to Python cognee.modules.users.models.User.

Enums

DataInput

MemoryEntry

Tagged union of typed memory payloads dispatched by remember_entry().

Traits

HasDataPoint

Read / write access to the embedded DataPoint of a typed container, plus a hook to recurse into nested DataPoint-bearing children.

Functions

classify_documents

Classify Data items into Documents based on file extension.

doc_type_for_extension

Map a file extension to a document type string, returning None for unrecognised extensions.

to_cognify_timestamp

Convert a raw LLM-extracted timestamp to a CognifyTimestamp with computed time_at. Returns None if the date is invalid (e.g. month=13).