AMAlexMercedAI

Alex Merced on agentic AI

AI should be open
to inspection.

I build open tools, specifications, and ideas for agentic systems that people can understand, govern, and move.

Explore the work ↓Read the knowledge base →
OPEN AGENTIC SYSTEM● LIVE
CONTRACTS
AGSOAPAAIS
↓ portable instructions
ROUTING
MERMerced AI
↓ governed delegation
EXECUTION
MagAgentLoroYour harness
MEMORYPOLICYRECORDS

WATCH / THREE ORIGINAL FILMS

Identity. Agents.
Open data.

Three scene-rich music videos connect the person, the agentic platform, and the open lakehouse underneath modern AI.

PERSON / MUSIC VIDEO

My Name Is Alex

A personal, music-driven introduction to the builder behind the work.

AGENTS / MUSIC VIDEO

Open the Stack

Composable agents, inspectable authority, model choice, and portable execution.

DATA / MUSIC VIDEO

The Iceberg Open Lakehouse

The open data foundation that makes governed analytics and AI possible.

THE THESIS

“The future of agentic AI is not one model, one agent, or one platform. It is replaceable components connected by open contracts.”

SELECTED OPEN WORK

Building the parts.
Defining the seams.

Projects spanning brokerage, execution, memory, governance, and interoperability. Each one has a full explainer in the knowledge base.

Agent brokerv0.5.1

Merced AI

A provider-neutral broker with portable profiles, durable sessions, and UI-presented runtime approvals.

Read the explainer →
Governed harnessv0.19.2

Loro

A governed harness with durable AAIS approvals across CLI, web, bots, and graph runs.

Read the explainer →
Developer harnessv1.1.1

MagAgent

A local-first agent harness with portable profiles, graph workflows, and reconnectable UI approvals.

Read the explainer →
Desktop workspacev1.0.0-rc.3

Mag Command Center

A cross-platform desktop cockpit with project context, graphs, profiles, operations, and native approval decisions.

Read the explainer →
Agent memoryv0.4.1

MagGraph

A graph-shaped memory layer for representing relationships, context, and retrieval paths.

Read the explainer →

OPEN CONTRACTS

Standards make ecosystems possible.

Open software is strongest when components share durable languages for work, identity, and human authorization.

Read about open contracts →
AGS
Agentic Graph Specification

Portable graph-shaped work with explicit tools, policy, budgets, and success criteria.

1.0 →OAP
Open Agent Profile

Portable agent identity, capabilities, authority, preferences, and state.

1.0 →AAIS
Agent Approval Interchange Specification

Exact, durable approval requests and decisions across CLI, web, desktop, and policy services.

1.0 →

DESIGN PRINCIPLES

Open is an operating model.

01

Portable by default

Agents, profiles, and workflows should move without being rebuilt around a single vendor.

02

Authority is explicit

An agent should know what it may do, what requires approval, and where its responsibility ends.

03

State is inspectable

Memory, context, and execution history should be understandable rather than trapped behind a magic interface.

04

Claims need evidence

Useful autonomy comes from traceable decisions, observable work, and verifiable outcomes.

THE AI BOOKSHELF / 31 TITLES

Ideas made
practical.

Explore Alex Merced’s nonfiction books on AI engineering, agentic systems, governance, evaluation, open models, semantic context, and the data foundations beneath modern AI.

Browse the complete book catalog ↗

Scroll to explore all 31 titles →

AM

ABOUT ALEX

Builder, educator,
open-systems advocate.

Alex Merced works across data infrastructure, developer education, and agentic AI. His focus is making complex systems legible, and giving builders open foundations they can adapt, audit, and own.

AlexMerced.com ↗OpenAgenticPlatform.com ↗Data work ↗Developer work ↗