MinasoftAI // Intelligence Architected

Reliable
Intelligence
for Industry.

Minasoft builds calm, reliable software for high-consequence operational environments. We curate logic for the physical world.

View The Thesis Documentation

System Live

Public systems are exposed through Minakeep, the MCP console, and inspectable GitHub repositories.

The Evolution

From Industrial Architecture to Scalable AI Operations.

Our journey began in the rigors of manufacturing and supply-chain architecture, evolving into the complex world of connected smart infrastructure.

Phase 01: Physical Origins

Industrial Foundations

Our foundation in manufacturing and supply-chain systems shaped our understanding of digital reliability and real-world operational failure modes.

The Curators

Hardware-coupled diagnostic systems that interpret complexity.

Voice And Assistant Systems

Atomic units and service layers designed for voice interfaces, AI assistants, and operator-facing workflows.

SDK And Platform Work

Real-time verification of operational intent versus execution across hardware-linked and distributed environments.

Bounded Inference

AI assistant actions remain inside strict safety and operational boundaries for the physical environment.

Proof Surface

Inspectable Engineering Proof.

Our engineering benchmarks are accessible via Minakeep and GitHub. We provide auditable logic, not black-box predictions.

PublicMinakeep Showroom

MCPProtocol Test Bench

GitHubRepository Evidence

AgentTooling Surface

Material_Ref: 014

Minakeep Terminal // Engineering_Verification

2026-05-15stevennana/minakeepPUBLIC

2026-05-15stevennana/mina-mock-mcpserverPUBLIC

2026-05-15stevennana/agent-toolingINSPECTABLE

// Linking public notes, protocol routes, and source repositories...

// Mock/protocol lab state is presented as test evidence, not production scale.

// Health: public endpoints reachable during verification.

Project Systems

Public AI systems operated by Minasoft.

Minasoft promotes working systems before claims. Minakeep and MCP Mock Server form the public product layer; the surrounding repositories show the agent-native tooling used to operate and extend them.

Minakeep

Public showroom

A self-hosted knowledge vault for private capture, selective publishing, and published knowledge that agents can treat as an MCP-readable source of truth.

Private capture
Selective publishing
Published notes for agent context

Live System GitHub Published Notes

MCP Mock Server

Public protocol lab

A public remote MCP test bench for repeatable protocol verification before production integrations.

Tool/resource/prompt permissions
Basic Auth and mock OAuth
Runtime library, inspector core, and CLI packages

Protocol Console GitHub Runtime Package Inspector Core MCP CLI

Supporting systems

Skills and bootstrap systems for agent-native development: wiki memory, Ralph loop scaffolding, REPL runtime, trend-to-Minakeep publishing, and documentation capture.

mina-wiki-skill mina-ralph-loop-bootstrap-nextjs mina-ralph-loop-bootstrap-python-cli mina-REPL-skill mina-github-trend mina-curls

View GitHub Repositories

Public Media Systems

A repeatable AI content pipeline, visible in public.

Minasoft also operates a public short-form media channel produced through AI-assisted image, prompt, and publishing workflows. It functions as an inspectable proof surface for recurring content operations.

Mina's Weekly Horoscope

Public short-form channel documenting AI-assisted visual content operations.

View Channel

Operational signals

Consistent visual system across weekly releases
Prompt-guided image generation with stable formatting
Repeatable publishing cadence on a live public channel

Capricorn Weekly

Stable layout, recurring frame logic, and consistent headline treatment.

Virgo Weekly

Template-driven image generation with repeatable color and type hierarchy.

Sagittarius Weekly

Public proof that the pipeline can sustain serialized topic variation.

Cancer Weekly

Character-led composition adapted without losing system consistency.

Scorpio Weekly

Evidence of lightweight AI media production beyond code artifacts alone.

Foundational principles for an era of industrial intelligence.

Architecting for reliability, privacy, and speed.

Bounded

Bounded Reasoning

Software must operate within defined physical constraints. Our frameworks ensure LLM outputs are converted into deterministic operational actions.

Architectural Detail_082

Request Technical Clearance.

Apply for credentials to access our sandbox environments and technical documentation.