dsh-mneme

🌏 简体中文 · English

---

🧬 Give Your LLM a Memory That Evolves

dsh-mneme is a cross-session memory plugin for DeepSeek Harness (DSH). It does more than "store" — it "manages": automatic background deduplication and merging, conflicts frozen for your review, a fully replayable audit trail, offline by default, and export to human-readable Markdown.

Mneme (Μνήμη) is derived from Mnemosyne, the Greek goddess of memory. She governs memory and dreams — just as autoDream quietly consolidates your memory store in the background.

What Problem Does It Solve

Every new chat feels like starting from scratch?

dsh-mneme gives DeepSeek Harness cross-session memory. Projects you've discussed, preferences you've mentioned, decisions you've made — the AI remembers them all, even after you close the window.

Scenario	Without Plugin	With Plugin
Discussed project requirements on Monday, continuing on Wednesday	"Could you describe your project again?"	"You mean the blog refactor we discussed last week? You mentioned wanting to use Astro."
Tell the AI your coding habits	Have to repeat yourself every session	Set once, persists indefinitely
Organize a large amount of material then close the window	Material is lost	Auto-archived, retrievable anytime

But what makes dsh-mneme trustworthy is what you can't see in the background. These are what truly distinguish it from "just a plugin that saves things."

Why You Can Trust It

• 🧾 Replayable & Auditable — Every automatic consolidation leaves a "decision receipt": input snapshot + decision details + result hash. The same consolidation can be replayed and reproduced, no silent errors, no untraceable changes.
• ⚖️ Conflicts Freeze First, Wait for Your Ruling (optional) — When two memories conflict, it doesn't make the call for you. Suspected conflicts are parked for review, compared side-by-side in the conflict queue on the status page, and resolved with one click (keep A / keep B / mark only). Changes take effect only after confirmation. For complex judgments, a human stays in the loop.
• 🔐 Memories Isolated by Agent & Workspace (optional) — With scopeEnabled, each memory is tagged with which agent wrote it and in which workspace; retrieval prioritizes the current session's scope (weighted hits, other scopes demoted but still visible); strictScope goes further — memories explicitly declared to belong to another scope are completely invisible in retrieval/injection (auto-tagged carriers are demoted only, not hard-blocked). Multiple agents and projects, zero cross-contamination.
• 🌙 Works Only When You're Idle (optional) — Automatically performs tiered archiving during idle time: frequently viewed items stay in the hot zone, long-unused items compress into summaries, stale items get fully archived. The memory store gets more refined with use, not bloated.
• 🧠 Local Semantic Search, Offline by Default — Comes with local Embedding and reranking built in; no API Key required; retrieval works even without a network.
• 📝 Two-Way Markdown Sync — Memories are local .md files you can open and edit anytime; manual edits are given priority and will not be overwritten by the system.
• 💾 Deleting a Chat ≠ Deleting Memories — Clearing the chat window keeps saved memories intact (configurable).

5-Minute Quick Start

# 安装插件
dsh plugin --profile web add @modusensus/dsh-mneme
dsh web

Works out of the box. To see its value in 5 minutes:

1. Chat: Start a new conversation and tell the AI about your preferences or a current project (e.g., "I prefer 4-space indentation for code").
2. Wait: Close the window, then open a new conversation. If it remembers what you just discussed, the memory has been saved.
3. Tune: Go to "Settings → Memory Settings" and enable the three switches below as needed (see Quick Configuration).

Quick Configuration (Optional)

Need	Config Key	Default	How to Change
Fully offline operation	embedProvider	openai	Change to local
Preserve memories when deleting chats	sessionLifecycleEnabled	false	Change to true
Auto-extract structured entities	entityExtractionEnabled	false	Change to true
Memory isolation per Agent / per project	scopeEnabled (add strictScope for stronger isolation — hard isolation applies to explicit declarations only)	false	Change to true

All of these are configured in DSH Settings Panel → Memory Settings. See the Configuration section for the full reference.

The Memory Loop in One Diagram

  写入 ──► 质量过滤（无用信息先拦下）
    │
    ▼
  SQLite + 本地 Markdown 镜像
    │（空闲时）
    ├─ autoDream ：去重 / 合并 / 归档 / 修正 / 冲突冻结
    └─ Sleep Mode ：分层压缩 + 模式发现 + 关系补全（可关）
    │
    ▼
  召回（混合检索 + 精排）──► 注入会话上下文

UI Preview

The panel is bilingual and follows your DSH interface language. The screenshots below are in Chinese; for the English version, see the English section at the end.

记录、浏览与筛选你的记忆。

自动从记忆里提炼实体，构建带属性的关系图谱。

状态面板一眼看清向量索引、LLM 消耗与自动巩固记录。

检索、实体抽取与记忆巩固开关都在设置里一站式配置。

可选写保护 Token，以及本地化的反馈通道，让记忆库完全本地、可审计。

Privacy

• Data is stored only on your local machine, never uploaded to any server
• Memories are Markdown files — human-readable and manually editable
• Zero network dependency by default; no API Key required
• No telemetry, no analytics, no remote logging

Docs

Document	Path
Full plugin docs (features / install / config / architecture)	dsh-mneme/README.md
Entity structure design	dsh-mneme/docs/ENTITIES.md
Semantic architecture	dsh-mneme/docs/SEMANTIC.md
Local model deployment guide	dsh-mneme/docs/LOCAL_MODEL.md
v0.3 → v0.4 migration notes (Sleep Mode)	dsh-mneme/docs/MIGRATION.md
Changelog	dsh-mneme/CHANGELOG.md
Security policy	SECURITY.md

🗺️ Roadmap

🧬 记忆基因 → 🛡️ 审计加固 → 💤 睡眠维护 → 🕸️ 召回融合与图谱 → ✨ 面板增强 → 🌡️ 自进化记忆 → 🕸️ 图谱增强

Version	Theme	Status
v0.3	Memory Genes: entities / attributes (with timeline) / relations	✅
v0.4	Sleep Mode: idle 4-phase deep maintenance	✅
v0.5	Recall fusion & memory visualization: BM25 + graph + hot memory	✅
v0.6	Session lifecycle: deleting a chat ≠ deleting memories	✅
v0.7	Self-evolving memory: heat decay + dual sleep protection + desktop workbench/feature toggles	✅
v0.8	Scope isolation (agent/workspace dual-dimension isolation + retrieval weighting + opt-in hard filter) + conflict queue manual arbitration + explicit attribution + ecosystem (stdio MCP server / graph recall axis / cold start / injection truncation & status bar / distillation reliability)	✅ Released (up to v0.8.4)

Full per-minor-version roadmap in dsh-mneme/README.md.

🧪 Local Development

cd dsh-mneme && npm install
npm test        # 1176 个测试
npm run stress  # 三轴线压测
npm run sync    # src → lib 同步

---

---

🙏 Acknowledgements

autoDream concept provenance (ideas credited, implementation original):

• Auto Dream in Claude Code (Anthropic, Memory 2.0): the conceptual origin — a background sub-agent consolidates memory files between sessions (deduplication, contradiction repair, decay pruning).
• Sleep-time Compute: Beyond Inference Scaling at Test-time (UC Berkeley & Letta, arXiv:2504.13171): the academic thread behind the "offline consolidation" idea.
• cc-haha: one of the early implementation references.

Building on the above work, dsh-mneme adds its own engineering developments: K-Means++ cluster pre-grouping, typed decision lists (keep / merge / archive / conflict / update, plus sleep-side supersede / differentiate) and a replayable sha256 digest audit chain (dream_runs / receipt_chain). If any source of inspiration is missing, feel free to open an issue.

🧬 Give Your LLM a Memory That Evolves

dsh-mneme is a cross-session memory plugin for DeepSeek Harness (DSH). It does not just store your memories — it manages them: background deduplication and merging, conflicts frozen for your review, a fully replayable audit trail, offline by default, and export to human-readable Markdown.

Mneme (Μνήμη) comes from Mnemosyne, the Greek goddess of memory and dreams — just as autoDream quietly consolidates your memory store in the background.

What problem does it solve

Every time you start a new chat, the AI acts like it's never met you?

dsh-mneme gives DeepSeek Harness cross-session memory. Projects you've discussed, preferences you've mentioned, decisions you've made — the AI remembers them even after you close the window.

Scenario	Without plugin	With plugin
Continue a project discussion from Monday on Wednesday	"Can you describe your project again?"	"You mean the blog refactor from last week? You mentioned wanting to use Astro."
Tell the AI your coding habits	Repeat every session	Set once, remember forever
Close window after organizing research	Notes are lost	Auto-archived, retrievable anytime

But what makes dsh-mneme trustworthy lives in the background you never see. These are the traits that set it apart from "a plugin that just saves things."

Why you can trust it

• 🧾 Replayable, accountable — every consolidation leaves a "decision receipt": input snapshot + decision detail + result hash. The same run reproduces the same outcome. No silent mis-merges, no untraceable changes.
• ⚖️ Conflicts freeze, you decide (opt-in) — when two memories contradict, it does not take sides for you. The suspected conflict is parked for review, compared side-by-side in the status-page conflict queue, and resolved with one click (keep A / keep B / mark reviewed). On hard judgments, a human stays in the loop.
• 🔐 Memories isolated by agent & workspace (opt-in) — with scopeEnabled, every memory is stamped with which agent wrote it and in which workspace; retrieval favors the current scope (weighted hits, out-of-scope demoted but visible); strictScope goes further — memories explicitly scoped to other agents/workspaces become invisible to search and injection (auto carrier labels are demoted only, never hard-blocked). Multiple agents and projects, zero cross-talk.
• 🌙 It works while you sleep (opt-in) — idle time triggers tiered archiving: frequent memories stay hot, stale ones compress to summaries, old ones archive. The store stays lean as it grows.
• 🧠 Local semantic search, offline by default — built-in local Embedding + reranking. No API key required; retrieval still works without a network.
• 📝 Two-way Markdown sync — memories are local .md files you can open and edit; human edits are respected, never clobbered by the machine.
• 💾 Delete the session ≠ delete the memory — clearing a chat window keeps what was saved (configurable).

5-minute quickstart

# Install the plugin
dsh plugin --profile web add @modusensus/dsh-mneme
dsh web

It works out of the box. To feel its value in five minutes:

1. Chat — start a session and tell the AI something about your preferences or a project (e.g. "I prefer 4-space indentation.").
2. Verify — close the window, open a new one. If it recalls what you said, the memory has landed.
3. Tune — open Settings → Memory Settings and flip the three switches below as needed.

Quick config (optional)

Need	Config key	Default	Change
Fully offline	embedProvider	openai	Change to local
Keep memories when deleting sessions	sessionLifecycleEnabled	false	Change to true
Structured entity extraction	entityExtractionEnabled	false	Change to true
Memory isolation per agent / workspace	scopeEnabled (add strictScope for stronger isolation — hard blocking applies to explicit declarations only)	false	Change to true

All of these live in DSH Settings → Memory Settings. Full config docs in the Configuration section (Chinese, bilingual file).

The memory loop in one diagram

  write ──► quality filter (drop noise first)
    │
    ▼
  SQLite + local Markdown mirror
    │ (when idle)
    ├─ autoDream   : dedupe / merge / archive / fix / freeze-conflict
    └─ Sleep Mode  : tiered compression + pattern discovery + relation completion (opt-in)
    │
    ▼
  recall (hybrid search + rerank) ──► inject into the conversation

Screenshots

The panel is bilingual and follows your DSH interface language. English shots below; see the Chinese section for the localized UI.

Record, browse and filter your memories.

Entities are extracted from your memories, building a relation graph with attributes.

The status panel shows your vector index, LLM spend and consolidation activity at a glance.

Retrieval, entity extraction and consolidation toggles are all configured in one place.

Optional write-protect token and feedback channels for a fully local, auditable setup.

Privacy

• Data stays on your machine only, never uploaded
• Memories are Markdown files, human-readable and editable
• Zero network dependency by default, no API key required
• No telemetry, no analytics, no remote logging

Docs

Doc	Path
Full plugin docs (features / install / config / architecture)	dsh-mneme/README.md (Chinese)
Entity structure design	dsh-mneme/docs/ENTITIES.md
Semantic architecture	dsh-mneme/docs/SEMANTIC.md
Local model guide	dsh-mneme/docs/LOCAL_MODEL.md
v0.3 → v0.4 migration (Sleep Mode)	dsh-mneme/docs/MIGRATION.md
Changelog	dsh-mneme/CHANGELOG.md
Security	SECURITY.md

🗺️ Roadmap

🧬 Gene → 🛡️ Audit hardening → 💤 Sleep maintenance → 🕸️ Recall fusion & graph → ✨ Panel enhancement → 🌡️ Self-evolving memory → 🔐 Scope isolation

Version	Theme	Status
v0.3	Gene: entities / time-boxed attributes / relations	✅
v0.4	Sleep Mode: idle 4-phase deep maintenance	✅
v0.5	Recall fusion & visualization: BM25 + graph + hot memory	✅
v0.6	Session lifecycle: delete session ≠ delete memory	✅
v0.7	Self-evolving memory: heat decay + sleep dual-protection + desktop workbench/feature toggles	✅
v0.8	Scope isolation (agent/workspace stamping + retrieval weighting + opt-in hard filter) + conflict review queue + explicit attribution + ecosystem (stdio MCP server / graph recall axis / cold-start bootstrap / injection truncation & status bar / distill reliability)	✅ Released (up to v0.8.4)

Full per-minor-version roadmap in dsh-mneme/README.md.

🧪 Local development

cd dsh-mneme && npm install
npm test        # 1176 tests
npm run stress  # three-axis stress test
npm run sync    # src → lib sync

---

Acknowledgements

Provenance of the autoDream concept (ideas credited, implementation original):

• Auto Dream in Claude Code (Anthropic, Memory 2.0): the conceptual origin — a background sub-agent consolidates memory files between sessions (dedupe, resolve contradictions, prune decay).
• Sleep-time Compute: Beyond Inference Scaling at Test-time (UC Berkeley & Letta, arXiv:2504.13171): the academic thread behind the offline-consolidation idea.
• cc-haha: an early reference for the implementation approach.

On top of these, dsh-mneme adds its own engineering: K-Means++ cluster pre-grouping, a typed decision list (keep / merge / archive / conflict / update, plus supersede / differentiate on the sleep side), and a replayable sha256-digest audit chain (dream_runs / receipt_chain). If any source of inspiration is missing, please open an issue.

📜 License

MIT