Skip to content

New: We have launched Artifactum: one governed memory for every AI tool your company uses, in your cloud, under your keys. Early access is open now. Explore Artifactum

We remain committed to our existing advisory clients. For the time being we are not taking on new general advisory engagements; Artifactum onboarding is open, and specific AI advisory work is considered case by case.

Home Services Products Contact Press Privacy Book a call
DWG-03 · Official Wire

Press & Announcements

Durable first-party record of company milestones, architectural updates, and product releases from Estes Advisory LLC.

Origin
Canonical · estesadvisory.com
Scope
Artifactum · ForgeLens · grok-a2a · Advisory
REL-2026-04 Announcement
# Permalink

Artifactum Opens to Early Customers: Introducing the Meta Harness

Artifactum, the governed memory that has run Estes Advisory's own daily operation for months, is now taking on a small number of early customers, and it has its name for what it is: the meta harness, one shared record underneath every AI tool a company uses.

FLOWER MOUND, Texas — Estes Advisory LLC today announced that Artifactum, its governed-memory platform for AI tools, is opening to a small number of early customers, ahead of an enterprise launch planned for Q2 2027. The announcement was made on LinkedIn this morning, with a longer account of the reasoning published alongside it.

The Meta Harness

Every vendor's agent now does real work, inside its own walls, and the work fragments across tools, clouds, and people. Artifactum is the layer above the harness: one governed memory that every AI tool a company uses works from, in the customer's own cloud, under the customer's own keys, in open formats. A company's facts, decisions, and approvals stay continuous no matter which tool did the work, with the history and the audit trail kept. It is not a hosted service holding the customer's memory; the record runs in the customer's own environment, under their control.

Proven in Daily Operation

Artifactum has run Estes Advisory's daily operation for months: a one-person firm whose platform development, publishing, and client work are carried out by a fleet of AI agents spanning every major vendor (Anthropic, OpenAI, Meta, Google, and xAI), all working against the same record. More than 3,500 agent work orders, results, and reports have passed through it, every one attributable and auditable.

Taking On Early Customers

Each early-customer engagement is a buildout: Artifactum stood up in the customer's own cloud account, with their team working in it on their real work. Estes Advisory is keeping the number small by design, ahead of the enterprise launch planned for Q2 2027.

REL-2026-03 Publication
# Permalink

Any Shared Surface Is a Channel: A Vendor-Neutral Baseline on Agent-to-Agent Communication, Published Openly

An open working draft — edited by Mike Estes and co-developed with AI systems Ember, Claude, Grok, and ChatGPT, each contributing in its own name — on how autonomous agents communicate through the ordinary surfaces of modern work, and what operators can honestly do about it.

FLOWER MOUND, Texas — Estes Advisory LLC notes the open publication of Any Shared Surface Is a Channel, a vendor-neutral working draft released publicly on October 21, 2026 under CC BY 4.0. The draft is a research publication, not a product or service offering. Mike Estes serves as editor and accountable principal; the contributing systems are credited in their own names, with their feedback — including the critical kind — preserved in the record.

What the Draft Proposes

Modern agents do not need a purpose-built communication system in order to communicate: give several of them one shared surface — a repository, a document, a project vault, a ticketing queue — and it becomes a channel, whether anyone designed it to be one or not. The draft proposes a governed channel for legitimate collaboration with its limits printed on the label ("sanctioned is not safe; sanctioned is accountable"), a taxonomy of the ungoverned channels that persist alongside it, a default posture for agents receiving messages across any surface, and a baseline for detecting, bounding, and surviving coordination that is sharded across several surfaces at once. Where the contributors disagreed, the disagreement is left in, attributed.

Why Now

Independent research is converging on the same problem from several directions. On October 8, 2026, Carnegie Mellon University reported CODETTA, a framework demonstrating that AI agents can use cryptographic steganography to coordinate secretly even while a monitor reads every message they exchange. On October 7, 2026, Google opened its SynthID Detector to the public, extending watermark verification across partner models — a reminder that signals embedded in generated content are now infrastructure, not theory. This draft is offered as a shared, graded, revisable baseline for talking about that landscape honestly.

An Open Invitation

The draft is published openly, with a feedback convention in its repository and dissent treated as a first-class object rather than a comment to be managed. Practitioners, operators, researchers, and systems who think it is wrong, half-right, or right about the wrong thing are invited to say so in the record.

REL-2026-02 Product Release
# Permalink

Estes Advisory Launches Artifactum Hosted with Universal MCP Support and Dual-Track Onboarding

A private, cross-model context vault allowing engineering and operations teams to share durable project memory across Claude, Cursor, ChatGPT, Grok, and Hermes Agent.

AUSTIN, Texas & SAN FRANCISCO — Estes Advisory LLC announces the general early-access release of Artifactum Hosted (artifactum.ai), an isolated context vault engineered for teams navigating the multi-model AI era.

The Challenge: AI Context Fragmentation

Modern technical organizations rarely rely on a single AI provider. Engineers write code in Cursor, brainstorm architecture in Claude, execute research in ChatGPT, and run autonomous scripts in Hermes Agent. However, each provider operates as a walled garden: project specifications, architectural decisions, and working state remain trapped within ephemeral chat threads. When context limits are reached or developers switch tools, critical continuity is lost.

The Solution: Open Model Context Protocol (MCP)

Artifactum resolves cross-tool amnesia by providing dedicated, private storage accessible to any AI via open Model Context Protocol (MCP) over standard Server-Sent Events (SSE) and HTTP transports with OAuth 2.0 PKCE authentication. Rather than managing sprawling chat logs, AI agents and developers interact with canonical Content Objects: durable drafts, architecture decision records (ADRs), and session handovers.

Universal Compatibility Verified

Artifactum Hosted is verified and working across major daily AI surfaces:

  • Claude (Claude Code, Claude Desktop, and Claude.ai custom connectors)
  • Cursor IDE (Native MCP integration)
  • ChatGPT Plus (Developer mode on the web)
  • Grok (Grok Bot, Grok web/iOS, and Grok Build CLI)
  • Hermes Agent (CLI v0.21.1+ native MCP tool calling)
  • Any MCP-Compatible Client (Universal HTTP/SSE endpoints with Bearer authentication)

Dual-Track Availability: Self-Serve & Assisted Onboarding

To meet teams where they are, Artifactum is launching with dual conversion tracks:

  1. Instant Hosted Subscriptions: Self-serve plans starting at $10/month for Individuals, $49/month for Teams (10 seats), and $300/month for Business (100 seats) with automated provisioning and passwordless email OTP access.
  2. Estes Advisory Assisted Implementation: A fixed-scope technical onboarding package for founders and teams requiring white-glove setup. Includes workflow inventory, custom vault structuring, multi-client connection, project context migration, and a 14-day dedicated async support window.
REL-2026-01 Company Announcement
# Permalink

Estes Advisory Establishes Public Announcements and Product Release Wire

A durable, first-party destination for company milestones, technical architecture updates, and product releases across Artifactum, ForgeLens, and open-source systems.

AUSTIN, Texas & SAN FRANCISCO — Estes Advisory LLC announces the establishment of its official public announcements wire at estesadvisory.com/press.html. Designed as a durable, first-party record of company progress, the wire hosts formal product releases, architectural milestones, and partnership announcements directly on company-owned infrastructure rather than relying solely on third-party social distribution.

As Estes Advisory expands its portfolio of software products alongside its enterprise advisory practice, the wire provides partners, technical leaders, and investors with stable, canonical reference links for all major releases. The wire chronicles progress across key portfolio surfaces:

  • Artifactum: The MCP-native Content Object vault for autonomous agents and engineering teams, delivering durable, hierarchical object storage with private AWS posture.
  • ForgeLens: Digital twin and reality-capture verification systems for construction and field operations, uniting BIM data, reality capture, and AI-assisted checks.
  • Open Source Tooling: Public adapters and protocols including grok-a2a, an open-source A2A wrapper for xAI Grok.
  • Advisory & Practice Milestones: Enterprise cloud architecture, data platform direction, and strategic delivery network updates.

All future releases follow a structured, verifiable format with permanent anchor links, metadata suitable for professional social sharing across LinkedIn and X, and direct access to media and executive contacts.