Mi-Memory: A Lifecycle Memory Framework for Personal AI

2026-07-21

AI Summary

Current personal AI systems predominantly rely on dialogue-level caching for memory, which falls short in ensuring state continuity, governability, and privacy compliance across multi-device and multimodal scenarios. To address these limitations, this work proposes Mi-Memory—the first personal AI memory framework supporting full lifecycle governance. Mi-Memory enables auditable, rollback-capable, and deployment-aware memory management through structured evidence payloads, diagnostic trajectories, policy artifacts, and gated records. Its core components—including MemStack, MemSense/MemFuse, D²ACCI/E²MEND, and LiteMem—facilitate evidence traceability, explicit policy evolution, and error localization. Evaluated on the LoCoMo, PersonaMem-V2, and LongMemEval benchmarks, the framework achieves structured memory accuracy rates of 93.59%, 57.24%, and 87.47%, respectively, demonstrating both its effectiveness and practical deployability.

Abstract

Personal AI is moving beyond chat-only interaction toward continuous services that span phones, cars, homes, wearables, cameras, and tools. In this setting, memory cannot remain a cache of prior conversations. It should serve as a continuity and governance substrate: preserving durable user state, grounding answers in multimodal and device evidence, supporting correction and forgetting, bounding policy evolution, and remaining deployable under latency, cost, privacy, and edge-cloud constraints. This technical report presents Mi-Memory, a lifecycle memory framework for Personal AI organized around four roles: Structure, Expansion, Evolution, and Deployment. A shared audit contract links these roles through four recurring artifact families: typed evidence payloads preserve source identity and provenance, diagnostic traces localize evidence loss across the serving pipeline, strategy artifacts make memory-policy changes explicit, and gate/rollback records bound accepted evolution. MiMemory instantiates the roles through MemStack, MemSense/MemFuse, D$^{2}$ACCI/E$^{2}$MEND, and LiteMem. In controlled-reference Structure evaluations, MemStack reaches 93.59%, 57.24%, and 87.47% on LoCoMo, PersonaMem-V2, and LongMemEval, respectively; other tracks report module-level, preliminary/internal, transfer-feasibility, or design-only evidence with explicit boundaries. MiMemory is a step toward auditable, evidence-gated, and deployment-aware memory systems for Personal AI. Project homepage: https://darwin-agent.github.io/Mi-Memory/ .

Problem

• Personal AI
• Memory Framework
• Multimodal Evidence
• Privacy Constraints
• Lifecycle Management

Innovation

• lifecycle memory
• evidence-gated memory
• auditable AI
• multimodal memory
• edge-cloud deployment