sgme
Run the following command in DeepSeek Harness:
dsh plugin install freehul/sgme
Paste the following prompt into your AI chat to install this plugin:
To install the sgme plugin in DeepSeek Harness, use the source repository https://github.com/freehul/sgme and run dsh plugin install freehul/sgme.
About this plugin
Most AI assistants treat every session as a first meeting: your preferences, ongoing projects, and character details have to be explained again, especially when you switch devices, models, or agents. SGME addresses this by turning memory into a persistent, searchable, and traceable layer that sits behind your AI conversations.
Its workflow is built around capture, distillation, and scenario-aware injection. Raw conversations are preserved as traceable records, then distilled into tagged memories such as facts, preferences, project states, and decisions, with deduplication, merging, and contradiction detection along the way. When a new conversation starts, SGME does not dump the entire memory store into context; it retrieves only what the current scenario needs, whether that is identity and recent status for casual chat, tech stack and pitfalls for coding, or plans and progress for work. It also supports proactive care signals, persona insight, unified search, shared knowledge bases, and skill management, so AI can remember you and also reach out when it matters.
SGME is for people who treat AI as a long-term companion, personal assistant, coding partner, or roleplay character, as well as for users working with multiple agents. Because it runs as a lightweight self-hosted engine with local storage and Chinese-first retrieval tuning, it suits users who care about privacy, continuity across sessions, and transparent, auditable memory.
Screenshots
Use Cases
- Use AI as a long-term assistant that remembers preferences and project progress across sessions.
- Automatically recall tech stacks, pitfalls, and coding habits during development.
- Let multiple AI agents share one memory brain and care proactively.
Best For
- Users who frequently work with multiple AI agents or DeepSeek Harness.
- Users who care about privacy and want memory data kept locally.
- Users who need AI to remember Chinese conversations, preferences, and long-term context.
Related Plugins
Traceable, searchable cross-session memory for AI agents that turns conversation knowledge into a typed knowledge graph and recalls relevant subgraphs instead of replaying full history, natively integrated with DeepSeek Harness.
Gives DSH AI cross-session long-term memory, to-do and skill management, plus multi-session orchestration, external AI delegation, and an infinite canvas that grows with you.
dsh-mnemon is a three-tier, pluggable, Agent-driven memory system for DeepSeek Harness, combining Runtime memory, Project Documents, and replaceable Memory Spaces with nine long-term providers.
An opinionated, zero-infrastructure file-based memory protocol: human-readable markdown files plus a bounded always-loaded index, curation discipline, and no database or embeddings.