dsh-mnemosyne-memory
Run the following command in DeepSeek Harness:
dsh plugin install fjzzwxp/dsh-mnemosyne-memory
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install fjzzwxp/dsh-mnemosyne-memory in your DSH terminal to install this plugin; full source: https://github.com/fjzzwxp/dsh-mnemosyne-memory
About this plugin
Every DSH session starts amnesiac—project conventions, architecture decisions, and last week's discussion all evaporate the moment the window closes. Mnemosyne plugs that gap: persistent, cross-session memory that survives restarts, so your agent actually remembers the project.
Under the hood it mirrors how human recall works. Vector-based semantic search turns a plain-language query into a precise hit on the right historical snippet. An LLM reflection pass fires every few turns, distilling decisions, insights, and team conventions into structured memory. Knowledge pages are auto-generated mid-session and then incrementally refreshed—no full rewrites. A codebase survey indexes 30+ configuration patterns in one pass. Multi-workspace isolation, Git-history seeding, and cross-session import all ship out of the box.
Whether you need fully offline operation with a local Ollama embedding model or prefer the higher accuracy of the free Gemini API tier, Mnemosyne is 100% free under the MIT license. Drop it into DSH and let every new conversation stand on the shoulders of every one before it.
Use Cases
- Keep project conventions and architecture decisions across DSH sessions
- Run fully offline with local Ollama for privacy-sensitive environments
- Share team knowledge with incremental page refreshes only
Best For
- Developers building AI agents on DeepSeek Harness
- Privacy-sensitive teams needing offline, zero-cost memory
- Individuals and teams who want agents to inherit past session knowledge
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.