Cross-Session Memory

aka Persistent User Memory, Long-Lived User Profile, Beat Agent Amnesia, No-Forget Memory, Agent Forgets Between Sessions, Session-to-Session Memory

category: memory · status: mature

Intent

Persist user-specific facts, preferences, and prior context across all sessions, threads, and devices.

Context

User-facing assistants where users expect continuity beyond a single conversation.

Problem

Per-thread memory loses everything between sessions; users repeat themselves; the assistant feels amnesic.

Forces

What to remember vs forget; user agency.
Privacy, deletion, portability requirements.
Cost of always-on memory loading per turn.

Solution

Maintain a per-user store of distilled facts (preferences, prior context, names, projects). Load relevant slices into each session's context. Provide explicit add/forget tools. Audit and surface memory entries to the user. Deletion controls and a user-visible memory inspector (delete / disable / export) satisfy regulatory and trust requirements.

Variants

Profile facts (key-value)

Distil each session into a small set of stable user facts (preferences, goals, constraints). Inject the facts into every new session's system prompt.

Distinguishing factor: structured key-value distillation

When to use: User-facing assistants where personalisation is the point and facts are short and stable.

Per-session narrative summary

Each session is summarised into a paragraph; recent summaries are retrieved and prepended to the new session's context.

Distinguishing factor: narrative summaries

When to use: Conversational continuity matters more than fact extraction; users expect the agent to 'remember the conversation'.

Embedded session retrieval

Past sessions are embedded and stored in a vector index; each new turn retrieves relevant past content by similarity.

Distinguishing factor: retrieval-time relevance

When to use: Memory is large and only a small fraction is relevant to any one new turn.

Tiered cascade

Combines profile facts, episodic summaries, and vector memory at different tiers; each tier retrieves at its own rate.

Distinguishing factor: multi-tier composition

When to use: Long-running personalised agents where no single memory shape is enough.

Constrains

Memory entries must be added through declared tools; the model cannot silently mutate persistent user state.

Consequences

Benefits

Continuity across sessions and devices.
Compounding usefulness over time.

Liabilities

Privacy obligations.
Memory hallucinations are stickier than chat hallucinations.

Known Uses

ChatGPT Memory
Claude Projects + memory
Letta
Lindy memory

Related Patterns

complements → short-term-memory
alternative-to → memgpt-paging
complements → session-isolation

References

OpenAI: Memory and new controls for ChatGPT (2024) blog