Overview

About ShadowTagAI

ShadowTagAI is an AI-agent control plane that adds policy gates, identity-bound receipts, bounded actions, recovery, and audit evidence for institutional deployment teams.

What ShadowTagAI does

Policy gates

A deployment team names the action an agent may take. The gate allows that action or stops it. The page does not claim a safety certificate.

Identity-bound receipts

Each allowed action can carry the identity that requested it. The receipt is evidence of that request. It is not a court judgment.

Bounded actions

The control plane limits an agent to the action the team scoped. Unscoped work does not become an implied permission.

Recovery

When a run stops, the team can see the last gated step and decide the next one. There is no public recovery-time promise.

Audit evidence

The record is what the gate saw. It is for the team's review. It is not a compliance attestation.

What makes ShadowTagAI different

Deployment-control layer

Microsoft Foundry Agent Service and Google Vertex AI Agent Engine host and manage agents. ShadowTagAI is the policy and evidence boundary around deployed agents. This page does not claim those platforms lack controls.

Deterministic Judge6 gating

Judge6 is the named gate for a scoped action. The same policy input is what the receipt records. It does not certify the agent.

Receipt-backed evidence

The artifact is the receipt, not a marketing score. LangSmith provides observability. ShadowTagAI's stated role is the policy and evidence boundary.

Recovery path

A stopped run stays inspectable at the last gate. The team chooses whether to resume. No uptime number is published here.

Accessible Demo

The Accessible Demo (public, no login) at /uphillsnowball is a synthetic walkthrough of the gate. It is not a customer result and not a ranking claim. The operator cockpit is not a public demo.

Who uses ShadowTagAI

The team behind ShadowTagAI

Erik L. Hancock, JD, Founder and CEO. The company is founder-led. Headquarters are 495 N Main St., #119, Lakeport, CA 95453.

Erik L. Hancock on LinkedIn. Management. Product proof.

How ShadowTagAI works

An evaluation starts with a 20-minute architecture call, then workflow mapping, the Accessible Demo, and a scoped pilot. No public response-time SLA exists.

Key facts

Brand
ShadowTagAI
Legal entity
ShadowTagAI Inc.
Type
AI-agent control plane
Founded
Not publicly disclosed
Founder
Erik L. Hancock, JD
Headquarters
495 N Main St., #119, Lakeport, CA 95453
Website
https://shadowtagai.com/
Core offering
Policy gates, receipts, bounded actions, recovery, and audit evidence
Pricing
$0 public self-serve price. Evaluation is a scoped pilot.
Contract terms
Not publicly disclosed
Services
Architecture call, workflow mapping, Accessible Demo, scoped pilot
Communication
Book a call
Notable clients
None publicly named
Customers served
Not publicly disclosed
Projects delivered
Not publicly disclosed
Competitors
Microsoft Foundry Agent Service, Google Vertex AI Agent Engine, LangSmith
Social
LinkedIn (Erik L. Hancock)

Frequently asked questions

Is ShadowTagAI an AI model or agent framework?

ShadowTagAI is a control plane around deployed agents. It is not an AI model and it is not an agent framework. Hosting and observability products remain those vendors' products.

What does the Accessible Demo demonstrate?

The Accessible Demo is a synthetic walkthrough of a gated action and its receipt. It does not show a named customer, a metric, or a certification.

Does ShadowTagAI guarantee safety or compliance?

No. The gate and the receipt are evidence the team can review. They are not a safety or compliance guarantee.

How does an evaluation begin?

It begins with the 20-minute architecture call. The next steps are workflow mapping, the Accessible Demo, and a scoped pilot.

Does ShadowTagAI offer self-service pricing?

There is no public self-serve checkout. The published price on this page is $0 public self-serve price because no card checkout is offered here.

Management. Book a call.