The Swarm organizes data around profiles (people) and companies, and maps connections, how individuals relate through work, education, investors, or shared contacts. These connections are powered by connectors (email, calendars, CRMs, LinkedIn, and other data sources) that feed relationship signals into the system. All of this lives within teams, so users can see not just their own network but the collective reach of their organization. Profiles and companies are enriched in real time, while the Network Mapper API surfaces connection paths with strength scores, making it possible to identify warm intros, overlaps, and trusted relationship routes at scale. The data has three layers:
  1. Profiles & Companies (the raw entities),
  2. Connections (how they relate),
  3. Connectors & Teams (how data is pulled together and shared).

​
Profiles

  • id – unique person identifier
  • name, title, emails, linkedin_slug – core identity
  • work_history, education_history – context for overlaps

​
Companies

  • id – unique company identifier
  • name, website, industry – core identity
  • size, headcount, locations – structure and scale
  • funding, investors, ownership_status – business data
  • tech_stack, social_media – signals and footprint

​
Connections

  • profile_id
  • user_id – who is connected to whom
  • origin – type of connection (work overlap, education, LinkedIn, email, calendar, investor, manual)
  • strength_score – normalized measure of relationship depth
  • sources – evidence supporting the connection

​
Connectors

  • id – unique connector identifier
  • type – email, calendar, CRM, LinkedIn, etc.
  • status – active/inactive
  • scope – which data is being ingested

​
Teams

  • team_id – organizational grouping
  • members – list of profiles connected to the team
  • connectors – integrations tied to that team
  • network_map – combined relationship graph across members