Independent software product company

We love what technology can make possible. We build it for real work.

Mythril Devs turns a deep interest in AI and modern infrastructure into focused products that make difficult systems more practical, understandable, and useful.

Built with curiosity. Operated with care.
Governed intelligence Durable context Operational evidence Endpoint trust Accessible web presence
Morgan Sinclair, founder of Mythril Devs

The person behind Mythril Devs

Morgan Sinclair

Founder · Infrastructure builder · Tucson, Arizona

For more than 25 years, Morgan Sinclair has built and operated technology in environments where failure has immediate consequences—from diagnostic imaging and managed business networks to 24/7 mining operations and acquisition-driven company growth. His work has crossed architecture, virtualization, systems, networks, identity, endpoints, cloud communications, observability, and governed enterprise AI.

Most recently, as the sole IT lead for a private-equity-backed home-services operator, Morgan rebuilt and scaled its technology environment from on-premises infrastructure to a cloud-first operation supporting more than 400 users and 1,400 devices across four locations in two states. He led cloud, identity, endpoint, network, multi-tenant, and acquisition integrations, then evaluated Microsoft 365 Copilot and deployed Claude for Teams with SSO, a responsible-use policy, and a cross-functional AI champion program.

Mythril Devs is not a departure from that work; it grew directly out of it. Every product began with a problem Morgan encountered personally—something that had to work reliably, securely, and without excuses. He builds to see how systems behave under real pressure, then turns what he learns into products for operators who have reached the same friction point and know there should be a better tool.

Follow Morgan’s public work on GitHub

What we’re building / 2026

Products born from problems we know firsthand.

Each began with real operational friction and advances only when the software proves useful in practice.

Live · invitation-only Alpha 01 / CURRENT FOCUS

Hoplon

Give your company AI that knows the business—and follows the rules.

A governed multi-model workspace and custom gateway for organizations that want practical AI without building an entire governance stack. Hoplon combines approved GPT, Claude, Gemini, and Grok access with policy enforcement, PII controls, budgets, cited research, and private source-grounded Projects.

Now

Gateway, workspace, onboarding, Projects, Compare, and provider cancellation are deployed. Access stays invitation-only while the operating model is validated.

HOPLON / CONTROL PLANE governed

Compare approved models against the same evidence.

MODEL / A 4 findings 2 cited sources
MODEL / B 3 findings 2 cited sources
Reviewed answer Save to Project Brief →
WorkersD1R2Queues VectorizeWorkers AI
Live · private Alpha 02

OWD

Owner-controlled memory for independent AI agents.

Bounded, revocable context plus durable handoffs, reviews, owner Decisions, and recovery—without making one agent the source of truth.

Next gate

A representative two-agent acceptance run.

Private beta · field validation next 03

AI Readiness

Map the work before automating it.

Structured intake, guided walkthrough capture, AI-assisted extraction, and human validation produce evidence-led operational maps.

Next gate

The first full real-company assessment-to-map run.

Phase 0 Alpha · pre-pilot 04

Goalie

A trustworthy local reflex layer for managed Windows endpoints.

Goalie turns narrowly verified profile-failure signals into privacy-safe, signed incidents without shipping raw Windows logs.

Next gate

Windows fixture validation, then the Intune pilot.

In development · not yet available 05

ChambaFluir

A repeatable web platform for Mexico’s smallest operators.

Curated sites, automated SEO hygiene, managed domain and email setup, WhatsApp-first contact, and metered Gemini notes planned through Cloudflare AI Gateway’s unified REST API—all behind a human accuracy gate.

Now

Product, integrations, pricing, and operating model are being built.

How we build

Powerful technology should still feel practical.

We enjoy working at the edge of what AI and modern infrastructure can do. Our job is to turn that potential into software people can understand, trust, and use—with clear interfaces, sensible limits, durable evidence, and human review where judgment matters.

Read our operating principles
  1. 01

    Bounded by design.

    Identity, policy, budgets, tenant scope, and revocation are product surfaces.

  2. 02

    Evidence over theater.

    Citations, provenance, recoverable state, and acceptance gates beat vague claims.

  3. 03

    Products, not projects.

    Repeatable software is owned, operated, and advanced by Mythril Devs.

  4. 04

    Humans keep judgment.

    Automation can assemble evidence; people approve consequential outcomes.

Why Cloudflare

A platform we genuinely enjoy building on.

Cloudflare lets a small product company move quickly without treating security, reach, or operational discipline as afterthoughts. Four product foundations already run on it; ChambaFluir is being designed for it.

GLOBAL EDGE Cloudflare shared substrate
Workers D1 + R2 Durable Objects Queues Workers AI Access + limits
01 / SHIP NARROWLY

One deployable unit at a time.

Workers and Static Assets keep product surfaces close to policy and data without turning each Alpha into an infrastructure program.

02 / KEEP BOUNDARIES

State fits the problem.

D1, R2, KV, Durable Objects, and Queues give each product a deliberate persistence and isolation model.

03 / EXPAND AFTER PROOF

Capacity follows validation.

A shared platform lets a lean company reuse proven patterns while every product keeps its own acceptance gate and operating boundary.

Workers Pages Static Assets D1 R2 KV Durable Objects Queues Vectorize Workers AI Access Rate Limiting

How we report progress

Shipped is not the same as validated.

Every product names what exists today and the next gate that could still stop expansion. Roadmap work is never presented as current capability.

CONTROLLED TESTING

Alpha

A primary path works end to end. Boundaries and operator behavior are still being proven.

Hoplon · OWD · Goalie
FIELD VALIDATION

Beta

A bounded release is deployed. Real-world use and review must validate the product in practice.

AI Readiness
NOT AVAILABLE

In development

The product and economics are being built. Planned integrations stay future tense.

ChambaFluir

Start a conversation

Curious about what we’re building?

We’d like to hear from people working on these problems, testing one of the products, building in the Cloudflare ecosystem, or simply interested in the technology.