Anthropic Auto Dream vs Claude Code auto-memory
Anthropic Auto Dream vs Claude Code auto-memory: side-by-side comparison of two claude code memory mechanisms systems — architecture, taxonomy, license, pricing, MCP/A2A support, and direct edges.
Anthropic Auto Dream · Claude Code auto-memory
Cost & capability
| Anthropic Auto Dream | Claude Code auto-memory | |
|---|---|---|
| Capability band | competent | competent |
| Capability composite | 70 | 70 |
| Cost tier | mid | — |
| Use cases | Code Generation Focused, Long Running Session | Code Generation Focused, Long Running Session, Offline Capable |
Where they differ (8)
Rows where both sides have data and the values disagree — the shortlist of dimensions that actually distinguish these two systems.
| Anthropic Auto Dream | Claude Code auto-memory | |
|---|---|---|
| Use cases | Code Generation Focused, Long Running Session | Code Generation Focused, Long Running Session, Offline Capable |
| Type | REM-style sub-agent memory consolidation | Per-project file directory |
| Created | 2026-05 (research-preview rollout in Claude Code; server-side feature flag) | 2025-03 (Claude Code v1.0 launched Feb/Mar 2025; auto-memory feature in subsequent releases; v2.1.59+ required) |
| Backend storage | local filesystem (~/.claude/projects/<project>/memory/) | file (~/.claude/projects/<project>/MEMORY.md) |
| MCP | native (first-party — Claude Code built on MCP) | native (first-party) — Claude Code consumes MCP |
| OpenTelemetry | no — local CLI feature | searched not found |
| Optimised for | background memory consolidation + bounded MEMORY.md size + surgical updates | Claude-Code-native + MCP-portable memory |
| Anti-fit | not for users on Claude.ai web (CLI-only); not for environments where consolidation logic must be auditable | not for non-Claude-Code clients (some are MCP-portable) |
At a glance
| Anthropic Auto Dream | Claude Code auto-memory | |
|---|---|---|
| Section | Claude Code memory mechanisms | Claude Code memory mechanisms |
| Tier | T1 | T1 |
| Type | REM-style sub-agent memory consolidation | Per-project file directory |
| Created | 2026-05 (research-preview rollout in Claude Code; server-side feature flag) | 2025-03 (Claude Code v1.0 launched Feb/Mar 2025; auto-memory feature in subsequent releases; v2.1.59+ required) |
| Pricing | bundled with Claude Code (no separate fee) | — |
| Funding | Anthropic total $12.4B+ raised; $40B valuation | — |
| Backend storage | local filesystem (~/.claude/projects/<project>/memory/) | file (~/.claude/projects/<project>/MEMORY.md) |
| Deployment | Local CLI (Claude Code) — Claude.ai web not in scope | — |
| API surface | /dream slash command + /memory toggle (Auto-dream: on) | — |
| Multi-tenancy | — | hard-isolation (per project) |
| MCP | native (first-party — Claude Code built on MCP) | native (first-party) — Claude Code consumes MCP |
| A2A | not supported | not supported |
| OpenTelemetry | no — local CLI feature | searched not found |
| Optimised for | background memory consolidation + bounded MEMORY.md size + surgical updates | Claude-Code-native + MCP-portable memory |
| Anti-fit | not for users on Claude.ai web (CLI-only); not for environments where consolidation logic must be auditable | not for non-Claude-Code clients (some are MCP-portable) |
Taxonomy
| Axis | Anthropic Auto Dream | Claude Code auto-memory |
|---|---|---|
| storage | file | file |
| retrieval | extraction-pull | injection |
| persistence | long-term | cross-session |
| update | consolidation | append-only |
| unit | fact | fact |
| governance | user-controllable | user-controllable |
| conflict | llm-arbitrate | append |
Pros & cons
Anthropic Auto Dream
Pros: Bounded MEMORY.md size (200 lines) prevents context bloat; surgical updates preserve user-edited content; user-toggleable; complements rather than replaces CLAUDE.md.
Cons: Closed-source; opt-out mechanism not formally documented; consolidation logic opaque; no audit log for what was pruned/merged; only Claude Code (not Claude.ai web).
Claude Code auto-memory
Pros: Per-project automatic memory built into the runtime — no plugin install, no external store, no extra prompts.
Cons: Single-vendor lock-in to Claude Code; no inspect/edit UI for what's been auto-saved; opaque to other tools.