CRMy.ai — Agent Memory Infrastructure for Sales

The agent memory layer that ships with a CRM.

Structured, typed, versioned working memory for AI sales agents. Open protocols, on your infrastructure. One call assembles everything your agent needs before each action.

What is briefing_get?

One API call that assembles full contact context before every agent action — record data, account health, recent activities, open assignments, typed context (objections, preferences, competitive intel with confidence scores), and stale data warnings.

Key Features

Typed context entries — Objections, preferences, competitive intel, research notes — each with confidence scores and actor attribution.
Stale data warnings — Surfaces what to distrust before acting, with expiration tracking.
Agent-human handoffs — Open assignments visible immediately so agents and humans stay coordinated.
MCP-native — Works with Claude Code, GPT, and any MCP-compatible agent. Also available via CLI and REST.
PostgreSQL-backed — Self-hosted on your infrastructure. No vendor lock-in.
Apache 2.0 — Free forever, no usage limits.

How It Works

Add CRMy to your agent: claude mcp add crmy -- npx @crmy/cli mcp
Agent calls briefing_get before each action to get full context.
Agent acts with context — drafts emails, updates records, logs new memory.

Why Not Build Your Own?

Every AI sales team rebuilds the same memory infrastructure: schema design, versioning, staleness tracking, confidence scoring, actor attribution, context assembly. CRMy gives you production-ready memory infrastructure so you can focus on building the agent.