IAM.Mind

Purpose

IAM.Mind preserves verified decisions and helps retrieve them in a later work session. Connected Codex, Claude Code, Cursor and DSH clients use the same memory account while retaining record provenance. Connecting memory does not configure a model or its route; the model profile remains a client setting.

Connect and verify

1. Open agent connections, install the supplied connector and sign in to IAM.Identity with your account.
2. History access, automatic outcome capture and lesson sharing are separate choices. Private chats remain excluded from history.
3. Save a synthetic decision in one client, wait for confirmed delivery and retrieve it in another. Check the original text and source agent.

Queue, search and permissions

Queued confirms the local queue; Delivered confirms server delivery. Server persistence and actual retrieval in another client are separate checks. Staging and production use separate data. A different account or missing history permission can change search results.

MEMORY.md, CLAUDE.md and AGENTS.md files are not automatically imported. Memory capture does not train model weights. Record correction and deletion depend on granted permissions. Never put tokens or passwords in messages.

Detailed customer paths

• Retrieve knowledge and verify sources.
• Connect four clients and verify exchange.
• Explore graph relationships.
• Project FAQ.

2mind has a separate early-access path. Its public demo does not prove that your working IAM.Mind memory is connected.