Context

How Context works

One private layer between your life and your AI.

Context gathers the sources you choose, makes the whole useful, and gives agents only the access they need.

01

Personal context

Choose the sources that help Context understand your days: calendar, health, voice notes, plans, and the people or projects you care about. Context turns them into a focused feed—not another inbox.

A quiet morning can stay protected. A meal plan can fit the meetings ahead. Training can respond to last night’s sleep without losing sight of the week.

A ranked ‘what matters now’ feedCalendar-aware prioritiesAdaptive meal and workout plans
02

Conversations

Meetings can be recorded in the background or offline, then transcribed on device. Decisions, commitments, people, and dates become useful context instead of disappearing into another transcript archive.

A conversation is one source among many. It informs the work and the day; it does not take over the product.

Offline and background recordingOn-device transcriptionDecisions connected to people, calendar, and work
03

Agent workflows

Through MCP and scoped connectors, agents can read the context they need, add artifacts, and run scheduled workflows. A founder can wake up to completed research, a verified build, or a concise daily brief.

Context keeps the workflow legible: epics, issues, artifacts, and gates show what happened, what is next, and where human judgment is required.

Research → spec → build → verifyScheduled runs and daily briefingsExplicit approval before consequential actions
04

Ownership and control

Your phone holds the canonical context. You choose what each agent can see, what it can add, and which actions must return to you for approval.

Permissions can be narrowed or revoked. The AI tools may change; the understanding you have built does not have to start over with them.

Phone-resident canonical contextScoped, revocable permissionsHuman approval at meaningful gates

Give every AI the context to be genuinely useful.