calm
A personal AI · built to be owned

Owned, not rented.

Lumen remembers what you approve, works across Mac and iPhone, and acts through governed tools — with receipts for what it used, what it did, and what still needs your decision.

Local-first by design. Clearly identified cloud capabilities are used only when configured and authorized.

Mac + iPhoneNative applications with an authenticated local bridge
6,600+Reliability tests in the current full-system gate
26 / 26Required audit stages passing before deployment
5 tiersAuthority from read-only work through explicit approval
The product

A system, not another chat window.

Lumen combines persistent context, live tools, explicit authority, and a native presence into one personal operating system. Status labels separate what works today from what is still being built.

Available

Correctable memory

Lumen can reuse reviewed facts, preferences, and corrections with provenance. You can inspect, change, or delete what it remembers.

Available

Governed action

Research, documents, calendar work, communications, and business tools run through explicit authority, approval, and receipt boundaries.

Available

Grounded answers

Live tools, evidence checks, visible failure states, and one committed answer replace confident guesses and hidden retries.

Available

Mac + iPhone

An authenticated device bridge and explicit conversation continuation connect Lumen across the devices its owner controls.

Private preview

Living Matrix

A shared cognitive presence visualizes committed system activity without exposing or pretending to reveal hidden chain-of-thought.

In development

Meeting copilot

A consent-first, Mac-led meeting experience is being built. Live capture, speaker identity, and automatic follow-up are not presented as shipped.

Living Matrix

See the system become active.

The Living Matrix turns committed tool activity into a shared visual presence on Mac and iPhone. Research, memory, communication, and creation wake distinct regions inside one cognitive body.

It visualizes system events and outcomes — never private hidden reasoning or chain-of-thought.

Private preview
Trust and privacy

Local-first. Explicit at every boundary.

01Owned foundation

Core intelligence can run on hardware the owner controls, without turning the relationship into a seat license.

02Honest data paths

Local and configured cloud paths are distinguished. External providers are used only for capabilities the owner enables and authorizes.

03Earned authority

Lumen acts inside the authority it has earned and asks before higher-impact work. A tool name or model suggestion is never permission.

04Receipts, not claims

Actions, failures, approvals, and committed answers produce evidence that can be inspected instead of relying on confident narration.

Plain-language privacy position

Lumen is local-first, not magically offline. Configured model providers, synchronization services, email, SMS, voice, image, and web tools may receive the information required to perform an authorized request. The roadmap includes stronger encryption, retention, deletion, and per-turn privacy receipts.

Evidence

Build status, not mythology.

Shipped foundation

One committed answer

Visible answer bytes are tied to a single durable verdict artifact across ordinary and tool-driven turns.

Shipped foundation

Typed authority

Communications, authentication recovery, and unavailable-tool facts come from structured provenance rather than persuasive prose.

Deployment candidate

Transactional rollout

Mac and iPhone artifacts are built before mutation, identity-checked, accepted together, and rolled back on failed required checks.

Status snapshot: August 21, 2026. Device dogfood follows the next accepted deployment.

The story

Built by an electrician and his AI engineering team.

Lucas Thomas runs Bolt Electric in Jacksonville, Florida. He is not a software engineer. In April 2026, he began directing AI collaborators to build the personal system he believed should already exist.

Lumen now spans native Mac and iPhone applications, local model infrastructure, memory, business operations, communications, research, and a governed deployment pipeline. The work continues in public view: capability by capability, test by test, with limitations stated plainly.

Private demos and pilots

Come see what Lumen can actually do.

Investor conversations, pilot interest, technical collaboration, or curiosity are welcome. Messages are reviewed by Lucas; Lumen may help prepare a response under his approval.

lumen@talklumen.com