> ## Documentation Index
> Fetch the complete documentation index at: https://docs.orbitmy.co/llms.txt
> Use this file to discover all available pages before exploring further.

# How Orbit thinks

> Understand how Orbit thinks: one person with many identities, a living graph enriched from LinkedIn, and natural-language search across every surface.

# How Orbit thinks

Orbit is not another contact list. It is a **living map of people** — resolved to one person, scored by real interaction, searchable in language you already use.

## The loop

```mermaid theme={null}
flowchart LR
  Import --> Enrich
  Enrich --> Search
  Search --> Capture
  Capture --> Enrich
  Enrich --> Signals
  Signals --> Search
```

1. **Import** — pull people from apps you already use
2. **Enrich** — fill titles, companies, LinkedIn, history
3. **Search** — ask in natural language across *your* graph
4. **Capture** — add who you just met (voice, text, photo)
5. **Signals** — notice when the graph changes

## One person, many identities

The same human might appear as:

* a LinkedIn profile
* a Gmail contact
* a calendar attendee
* someone you mentioned in a note

Orbit merges those into a single **contact** with multiple identities underneath. Searching “Sarah at Stripe” should not return three half-empty cards.

## Relationship strength

Orbit records **interactions** — meetings, emails, messages, notes you capture — and uses them to understand closeness. That is why answers can sound like “you met at…”, not just “here is a LinkedIn URL.”

## Surfaces share one graph

Mac, web, iPhone, and Text Orbit read and write the **same** network. Capture on phone → search on Mac. Import on Mac → browse on phone.

<Tip>
  Day-1 tip: import the source with the densest graph first (usually LinkedIn on Mac). Search quality follows coverage.
</Tip>