dsh-meow-memory
Run the following command in DeepSeek Harness:
dsh plugin install Phant0Meow/dsh-meow-memory
Paste the following prompt into your AI chat to install this plugin:
To install this plugin in DeepSeek Harness, use the command dsh plugin install Phant0Meow/dsh-meow-memory. Full source code: https://github.com/Phant0Meow/dsh-meow-memory
About this plugin
Most LLM sessions start from scratch, forcing users to repeat context, preferences, and project state in every new conversation. dsh-meow-memory brings true cross-session memory to DeepSeek Harness by persisting workspace knowledge in a local SQLite database, so the AI automatically remembers who you are, where your project stands, and what you've agreed upon.
The plugin uses a seven-layer memory model covering the AI's own identity, user profile, project details, atomic facts, lessons, ongoing topics, and design rules. It injects a static "long-term memory block" before the first user message while keeping the system prompt cache-friendly, and from the second message onward, each user message triggers keyword-based recall that injects only the two most relevant memories. Tools for writing, searching, updating, and deduplicating memories are available, with project anchoring to avoid false matches during casual chat. During idle time, the plugin performs per-window "dream" consolidation, automatically refining and freezing the knowledge gained in that session.
It's ideal for anyone doing long-term projects, continuous writing, coding, or research inside DeepSeek Harness. With zero runtime dependencies and built-in Node sqlite, it installs cleanly; client-side touches like collapsed injection bars, reflection cards, and crescent moon icons keep memory working in the background without disrupting your workflow.
Use Cases
- Long-term project tracking and context continuity
- Personal knowledge base and preference accumulation
- Information recall across multiple sessions
Best For
- Developers and researchers who need persistent context
- Users engaged in long-form writing or creative work
- DeepSeek Harness users who want to avoid repeating background 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.