Anthropic Auto Dream vs codebase-memory-mcp (DeusData)
Anthropic Auto Dream vs codebase-memory-mcp (DeusData): side-by-side comparison of two claude code memory mechanisms systems — architecture, taxonomy, license, pricing, MCP/A2A support, and direct edges.
Anthropic Auto Dream · codebase-memory-mcp (DeusData)
Cost & capability
| Anthropic Auto Dream | codebase-memory-mcp (DeusData) | |
|---|---|---|
| Capability band | competent | competent |
| Capability composite | 70 | 45 |
| Cost tier | mid | free |
| $/Mtok input | — | 0 |
| $/Mtok output | — | 0 |
| Use cases | Code Generation Focused, Long Running Session | Code Generation Focused, Long Running Session, Offline Capable, Analytical Summarization |
Where they differ (14)
Rows where both sides have data and the values disagree — the shortlist of dimensions that actually distinguish these two systems.
| Anthropic Auto Dream | codebase-memory-mcp (DeusData) | |
|---|---|---|
| Capability composite | 70 | 45 |
| Cost tier | mid | free |
| Use cases | Code Generation Focused, Long Running Session | Code Generation Focused, Long Running Session, Offline Capable, Analytical Summarization |
| Type | REM-style sub-agent memory consolidation | Tree-sitter knowledge-graph MCP for code |
| Created | 2026-05 (research-preview rollout in Claude Code; server-side feature flag) | 2026 (active 2026 release cadence; latest v0.6.1 2026-05-04) |
| Pricing | bundled with Claude Code (no separate fee) | free / open source (MIT); no paid tier |
| Funding | Anthropic total $12.4B+ raised; $40B valuation | searched not found |
| Backend storage | local filesystem (~/.claude/projects/<project>/memory/) | SQLite (WAL mode) at ~/.cache/codebase-memory-mcp/ |
| Deployment | Local CLI (Claude Code) — Claude.ai web not in scope | Local — single static binary (macOS arm64/amd64; Linux arm64/amd64; Windows amd64) |
| API surface | /dream slash command + /memory toggle (Auto-dream: on) | MCP (stdio) — 14 tools incl. index_repository / search_graph / trace_call_path / query_graph (Cypher-like) / get_architecture / detect_changes / manage_adr / search_code / get_code_snippet / ingest_traces / list_projects / delete_project / index_status / get_graph_schema |
| MCP | native (first-party — Claude Code built on MCP) | native (first-party MCP server) |
| OpenTelemetry | no — local CLI feature | searched not found |
| Optimised for | background memory consolidation + bounded MEMORY.md size + surgical updates | structural code knowledge graphs (token-efficient; 99.2% reduction vs grep) + sub-ms queries + 155-language tree-sitter coverage |
| Anti-fit | not for users on Claude.ai web (CLI-only); not for environments where consolidation logic must be auditable | not for non-code memory; not for natural-language conversation memory; SEMANTICALLY_RELATED edges require optional embedding pipeline |
At a glance
| Anthropic Auto Dream | codebase-memory-mcp (DeusData) | |
|---|---|---|
| Section | Claude Code memory mechanisms | Claude Code memory mechanisms |
| Tier | T1 | T1 |
| Type | REM-style sub-agent memory consolidation | Tree-sitter knowledge-graph MCP for code |
| Created | 2026-05 (research-preview rollout in Claude Code; server-side feature flag) | 2026 (active 2026 release cadence; latest v0.6.1 2026-05-04) |
| Latest release | — | v0.6.1 / 2026-05-04 |
| License | — | MIT |
| GitHub | — | 2.1k★ (DeusData/codebase-memory-mcp) |
| Pricing | bundled with Claude Code (no separate fee) | free / open source (MIT); no paid tier |
| Funding | Anthropic total $12.4B+ raised; $40B valuation | searched not found |
| Backend storage | local filesystem (~/.claude/projects/<project>/memory/) | SQLite (WAL mode) at ~/.cache/codebase-memory-mcp/ |
| Deployment | Local CLI (Claude Code) — Claude.ai web not in scope | Local — single static binary (macOS arm64/amd64; Linux arm64/amd64; Windows amd64) |
| API surface | /dream slash command + /memory toggle (Auto-dream: on) | MCP (stdio) — 14 tools incl. index_repository / search_graph / trace_call_path / query_graph (Cypher-like) / get_architecture / detect_changes / manage_adr / search_code / get_code_snippet / ingest_traces / list_projects / delete_project / index_status / get_graph_schema |
| MCP | native (first-party — Claude Code built on MCP) | native (first-party MCP server) |
| A2A | not supported | not supported |
| OpenTelemetry | no — local CLI feature | searched not found |
| Optimised for | background memory consolidation + bounded MEMORY.md size + surgical updates | structural code knowledge graphs (token-efficient; 99.2% reduction vs grep) + sub-ms queries + 155-language tree-sitter coverage |
| Anti-fit | not for users on Claude.ai web (CLI-only); not for environments where consolidation logic must be auditable | not for non-code memory; not for natural-language conversation memory; SEMANTICALLY_RELATED edges require optional embedding pipeline |
Taxonomy
| Axis | Anthropic Auto Dream | codebase-memory-mcp (DeusData) |
|---|---|---|
| storage | file | graph |
| retrieval | extraction-pull | structural-query |
| persistence | long-term | project-bound |
| update | consolidation | replacement |
| unit | fact | fact |
| governance | user-controllable | user-controllable |
| conflict | llm-arbitrate | deterministic-overwrite |
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).
codebase-memory-mcp (DeusData)
Pros: Single-binary zero-deps deployment; 155-language tree-sitter coverage; structural queries replace grep at 120x fewer tokens; SLSA-3 + Sigstore + VirusTotal; team-shared graph artifact format; preprint with reproduction-friendly evaluation.
Cons: Preprint not peer-reviewed; founders/company opaque (DeusData has no disclosed principals); no managed/cloud option; SEMANTICALLY_RELATED edges depend on optional embedding pipeline.