The product

Memory that goes into the task,
and comes back for review.

Decision Memory is the approved decision layer for human–AI teams. It preserves the decisions, rules, skills, and authorizations that define how your organization works — and keeps a human in the loop for everything that becomes trusted.

Four governed objects

Not notes. Structured memory.

Live
Decision

What was chosen and why: rationale, evidence, rejected alternatives, scope, status, lineage.

Locally validated
Rule

What must / must not / requires approval. Versioned, linked to the decision behind it.

Locally validated
Skill

How a class of work should be done: patterns, playbooks, dependencies, write-back contracts.

Direction
Authorization

Which actor may do what, on what, under which conditions. Runtime enforcement stays separate.

// one task, four objects
Decision → use JSON for machine artifacts
Rule → never store them in Excel
Skill → define schema · validate · store example
Auth → agent may write repo;
     external upload needs approval
Governance

AI proposes. Humans approve.

Agent submits

Candidate

Agents retrieve permitted memory and submit proposals — never approving organizational memory.

Human

Review

A person reviews each candidate with its evidence and reasoning in context.

Outcome

Approve · edit · reject

Imported or generated rules stay candidates until reviewed.

Decision Memory is a memory & governance product. Your runtime still enforces actions.

For agents · MCP

Four tools your agents call.

dm.search_decisions

Search approved decisions.

dm.get_decision_context

Retrieve permitted decision context for a task.

dm.check_conflicts

Check possible conflicts across objects.

dm.propose_decision

Submit a decision to human review.

Try it

Test DM with one agent workflow.