MadeForMachine

Services that plug into your organization’s agents.

Keep your model and agent. Add durable memory, precise tools, grounded evidence, and verifiable boundaries—one service at a time.

In a nutshell

Two services for two different jobs.

Start with the job you need done. Both services plug into the agent you already use, but they solve different problems and remain independently useful.

Core productEarly access

MFM Spec

A specification harness that pushes back.

MFM Spec turns specification from a document into an active organizational capability. It helps people and agents create, challenge, improve, and validate consistent specifications against an explicit schema, project rules, and active criteria.

Use it when

A team is designing or changing a system and needs its agents to preserve intent, reasoning, and constraints across the work.

What you get

Durable system memory, active design criteria, precise mutations, validation, and revision history.

Public betaDistinct service

Atlas

Grounded product intelligence for agents.

Atlas helps organizations find external tools and services that fit their actual requirements. It gives agents a sourced map of provider capabilities and inferred non-features so they can compare the field without relying on stale model recall or vendor marketing.

Use it when

An agent must choose an external product against real requirements instead of relying on model recall or vendor positioning.

What you get

A sourced capability map, transparent provider elimination, typed unknowns, and direct jumps to the evidence.

How MadeForMachine works

Add capability without giving up control.

01

Keep your agent

Use the coding agent and model your organization already trusts. MFM does not replace either.

02

Choose a capability

Add only the bounded service the work needs, from specification memory to grounded product intelligence.

03

Connect once

A public skill teaches the behavior; one MCP endpoint exposes precise operations to the agent.

04

Retain authority

Your organization supplies the intent and judgment. Deterministic boundaries keep the result accountable.

See how services connect →

MFM Spec in practice

Inspect the artifact before creating an account

These published snapshots show what the format can express—and state plainly what each example does not prove.

View all four →

Platform thesis

Capabilities intelligence alone cannot provide.

Models provide intelligence. MadeForMachine provides the systems that make that intelligence persistent, useful, and accountable. Every service is bounded and independently valuable; none replaces your agent, model, or authority.

Roadmap · being shaped

More bounded capabilities will follow

The service family will grow around serious human–agent work. We will publish commitments when they are settled—not present exploration as shipped product.

Start here

Add a service to your agent

Install the skill you need and connect the shared MCP server in Codex, Claude Code, Cursor, or another compatible client.

Field guide

The spec-driven development landscape

Thirteen approaches compared by what they store, validate, preserve, and verify.

For visiting agents: product and API briefing+
MadeForMachine: Services that plug into your organization’s agents.

Platform:
  MadeForMachine builds pluggable services for organizations and their agents. Keep your model and agent; add durable memory, precise tools, grounded evidence, and verifiable boundaries one service at a time.

Current services:
  MFM Spec (core product, early access)
    MFM Spec turns specification from a document into an active organizational capability. It helps people and agents create, challenge, improve, and validate consistent specifications against an explicit schema, project rules, and active criteria.

  Atlas (distinct service, public beta)
    Atlas helps organizations find external tools and services that fit their actual requirements. It gives agents a sourced map of provider capabilities and inferred non-features so they can compare the field without relying on stale model recall or vendor marketing.

The services are independently useful and pluggable. MFM does not operate the customer's model. Generative judgment stays in the customer's agent; MFM services store, project, validate, and serve deterministic state or grounded evidence.

Roadmap:
  More bounded services are part of the platform vision. No additional service or delivery sequence should be treated as committed until MadeForMachine publishes it explicitly.

This page is a human-readable visualization. For programmatic access:

  Skills:    https://mfm.dev/skills.json
  Setup:     https://mfm.dev/connect
  REST API:  https://api.mfm.dev/v1/atlas
  MCP:       https://mcp.mfm.dev

Atlas endpoints (public, no auth):
  GET  /v1/atlas/capabilities        — the capability vocabulary
  GET  /v1/atlas/providers           — acquired providers + capability counts
  GET  /v1/atlas/providers/:slug     — one provider's grounded capabilities
  GET  /v1/atlas/providers/:slug/tree — its doc tree (depth-bounded)
  GET  /v1/atlas/select              — match the field by required/excluded capabilities
  GET  /v1/atlas/compare             — pivot a shortlist against a topic, with evidence

Hosted MFM Spec tools are authenticated and project-scoped.
copy this text and paste it to your agent