Back

Omnichannel Intelligence Demo

See how Graphiti transforms raw data into authoritative facts

Omnichannel Inputs

Raw data from various channels before temporal resolution

Mobile Network Event (Structured / Deterministic)

BSS Event

This is a factual system event. It may directly update user state.

Customer Support

Customer Support Interaction (Conversational / Interpretive)

Chat

This is what the user said. Facts must be inferred and validated.

Channel context used by Graphiti for temporal resolution (user_id, channel, timestamps). Auto-populated when loading presets.

Avatar Memory
Personal Context (Avatar Memory)

Personal Context (Conversational / Interpretive)

Avatar Memory

Conversations that reveal personal details - pets, relationships, emotions, interests, goals.

Pets

My cat Mochi...

Relationships

My mom, grandma...

Emotions

I'm stressed...

Interests

Playing Genshin...

Goals

I want to...

Personality

I'm introverted...

Graphiti Context Engine

Authoritative temporal facts derived from all channels

Current Authoritative State

SOURCE: GRAPHITI

Current user state after resolving all channels and history

No state resolved yet

Submit data to see the current state

Knowledge Graph

KNOWLEDGE GRAPH

Interactive visualization of entities and relationships

No graph data yet

Process data to visualize the knowledge graph

Temporal Fact History (Why This Is True)

TEMPORAL

Older facts are retained but not active (valid_to is set)

No temporal facts yet

Episodic Signals (What We Observed)

NON-AUTHORITATIVE

Signals derived from episodes - used for AI context but not stored as facts

No episodic signals yet

AI Context Projection

DERIVED

Recomputed on demand from Graphiti facts + episodic signals. Never stored - always fresh. This is what an LLM would receive as context.

No context projection yet

Submit data to generate AI context