Three blindnesses. Three ways to see.

Sava is the interpretive layer. Using Sava or a supported AI assistant, you can ask the question every project lead asks eventually: how do I know this person, really?

Before a meeting with a utility's interconnection team, you can ask Sava, or a supported AI assistant, who's in the room, who they trust, what was said 3 months ago, and why the quiet engineer still matters.

Answers come from what your Orbit holds and point back to where they came from, so you can check them before you act.

Surface

Plain-language questions. Answers point to their sources in Orbit.

Strengths

Pre-meeting orientation, routes to the right person, why-this-person-matters answers.

Pairs with

MCP. Claude, ChatGPT, Notion, and Grok read and write the same Orbit Sava does.

Orbit is a persistent, navigable map of the people, organizations, projects, sites, and events in your world. Each relationship holds context.

A person in Orbit carries a Context Signature: who introduced them, how strong the relationship is, when it was last touched, why it matters, what it's tied to.

The relationship between a developer and an AHJ has a different shape from the one between an owner's rep and a lender. Orbit keeps that shape.

Node types

People, organizations, projects, sites, jurisdictions, utilities, permits, events.

Relationship context

History, strength, obligation, timing, source.

Persistence

The map stays with the project as people change roles.

Last year below, the next 90 days approaching from up ahead, next year above, rendering long timelines legible.

Helix is designed so a person becomes a thread through the moments your team actually touched them, with scheduled meetings, promised deliverables, RFP windows, and comment deadlines approaching from ahead.

Today, Orbit holds dated events and approaching commitments on each relationship. Helix, the full temporal view, is emerging.

Kosmos turns plain language into people, organizations, relationships, and dated events. It's how a hallway conversation becomes something the team can find later.

Debrief after a site walk, paste in meeting notes, or speak a recap on the drive back.

Kosmos parses it and writes it into Orbit. No forms.

Input

Voice notes, pasted text, meeting transcripts, email threads, photographs of a business card or a whiteboard.

Output

People, organizations, relationships with context, dated events, and where each came from.

Edits

Anything Kosmos writes can be corrected in Orbit afterward.