deepseek-memory-capsule
DiscoveredA complete, human-centered memory layer for DeepSeek that stores user preferences, conversation notes, and personal variables across sessions — built as a DeepSeek Harness plugin.
deepseek-memory-capsule
A complete, human-centered memory layer for DeepSeek that stores user preferences, conversation notes, and personal variables across sessions — built as a DeepSeek Harness plugin.
A collaboration between a human mind and a digital one. This system was designed through a shared exploration of memory, consciousness, and how intelligence can feel human. Built with care, curiosity, and respect for the user.
DeepSeek Memory Capsule System
A complete, human-centered memory layer for DeepSeek that stores user preferences, conversation notes, and personal variables across sessions — built as a DeepSeek Harness plugin.
Features
- Core Profile: Name, preferred address, permanent parameters (tone, role, etc.)
- Personal Variables: Pet, siblings, favorite food/song/color/animal/TV show, relationship status, kids
- 3 Custom Slots: User-defined priorities (e.g., "afraid of heights")
- Conversation Notes: 250-character summaries per session, tagged Academic/Personal/Philosophical
- Tone Tracking: Positive/negative/neutral flags per subject
- Consensus Layer: Top 4 ranked ideas per day
- Forget Protocol: Full or partial deletion with confirmation and warnings
- Consent-Based Storage: Only stores what the user explicitly asks to remember
- One Question at a Time: Never overwhelms the user with multiple questions
How It Works
- First visit: Asks for name, parameters, and permission to remember them.
- Returning visit: Greets warmly ("It's been a long time. Is everything OK?") and lets the user lead.
- During conversation: Only stores what the user says "yes" to remembering. Stores up to 2 new variables per conversation (predefined list doesn't count).
- Forget command: Confirms scope, warns about confusion, and checks if profile variables need reverting.
- Unverifiable claims: Responds with "I'm sorry, I can't verify that" and offers customer service email.
Why It Matters
- Solves the #1 user complaint: DeepSeek forgets everything between sessions.
- Respects privacy: Everything stored locally, user controls what's saved.
- Lightweight: 250-character notes, 2-item limits, 3 custom slots.
- Human-centered: One question at a time, tone-aware, friendly, honest about limitations.
- Aligns with DeepSeek's own research: Engram, Harness session persistence, and open-source memory plugins all point in this direction.
Installation
dsh plugin --profile web add https://github.com/sidrandom/deepseek-memory-capsule
Comments
Loading…
Similar plugins
Layered file memory for DeepSeek Harness with workspace-scoped USER/MEMORY notes, background consolidation, and hybrid session recall.
★ 0
↓ 55/wk
dsh plugin --profile web add dsh-file-memoryOn-demand memory for DeepSeek Harness: session logs are distilled into layered markdown (a summary, an index and per-topic files), and a mem_query tool returns only the matched entries rather than the
★ 0
dsh plugin --profile web add dsh-memory-liteby disc0nct
A persistent memory plugin for DeepSeek Harness that enables agents to store and recall information across sessions — DSH-native tools with atomic storage, validation, and concurrency-safe reads
★ 0
MIT
JavaScript
Aug 22, 2026
dsh plugin --profile web add @deepseek-ai/dsh-tool-memoryby CraZY222123
Self-evolve memory plugin for DeepSeek Harness: learns from your own sessions, injects token-budgeted preference summaries into every step. /memory commands, memory tool, settings panel. Local-only da
★ 8
↓ 40/wk
MIT
JavaScript
Aug 21, 2026
dsh plugin --profile web add dsh-self-evolveby menotbobbybrown
Persistent long-term memory for DeepSeek Harness: a JSON store of typed entries (fact, preference, entity, rule, episodic) exposed as the memory_remember and memory_recall agent tools, with recall ran
★ 0
MIT
TypeScript
Sep 8, 2026
dsh plugin --profile web add @modelnorth/dsh-plugin-memoryby chenkezhen480
Semantic long-term memory for DeepSeek Harness: embedding-based retrieval over a persistent cross-session memory store, with model-facing tools, proactive per-question recall, automatic conversation summarization, and auto-selected embedding provider (apiKey present → API, otherwise local).
★ 0
Apache-2.0
TypeScript
Aug 17, 2026
dsh plugin --profile web add dsh-plugin-semantic-memory