Personal AI·AlphaPreview

IAM.Mind

Local-first AI workspace for project memory, citations, agents and runs.

Landing Application
IAM.Mind interface
Current public IAM.Mind surface. The screenshot was verified when the documentation was published.

Key capabilities

  • Local chat and owner-scoped memory
  • Citations, restart persistence and deletion
  • Chat, Agents, Knowledge, Runs and System surfaces

Role in the ecosystem

IAM.Mind is a personal workspace in which chat, knowledge, memory and launches belong to the owner of the device/workspace. Local-first architecture reduces dependence on shared corporate memory and gives the user control over saving and deleting context.

Product surfaces

  • Chat — local dialogue with the selected model profile;
  • Memory — allowed facts and context between sessions;
  • Knowledge - sources and citations;
  • Agents - repeatable instructions and capabilities;
  • Runs — execution history, statuses and artifacts;
  • System — models, storage and device status.

Memory and evidence

Memory retention is a separate action, and not hidden retraining of the weights. A record is associated with an owner, source, and time; user can check, fix and delete it. Knowledge Answers save quotes to the context could be checked outside the model.

Status and roadmap

AlphaPreview captures the proven local scope. Active promotion of Sleep, autonomous growth, C0, SkillCells/COF-1 and multimodality proven and ready for BetaPreview. Synchronization between devices and Marketplace/swarm commerce remain separate roadmap directions without promises of deadlines.

Limit of responsibility

IAM.Mind is not shared tenant memory and does not update model weights on the worker user content. For external early-access product it is used separate surface 2mind.