dsh-memory
Run the following command in DeepSeek Harness:
dsh plugin install ben7am1n/dsh-memory
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install ben7am1n/dsh-memory in your DeepSeek Harness terminal to install this plugin; the source code is available at https://github.com/ben7am1n/dsh-memory
About this plugin
The DeepSeek Harness extension cookbook long reserved a slot for memory-a prompt section plus a handful of tools-yet nothing ever shipped, so every new session starts from a blank slate and the user must re-explain preferences, project context, and past decisions. dsh-memory closes that gap with a single local SQLite file: no embedding service, no API key, no sidecar process. All data lives in one .db file and survives process restarts.
The plugin exposes three tools to the model: memory_write stores one self-contained durable fact (optionally tagged and pinned), memory_search performs keyword retrieval across text and tags, and memory_forget removes an entry that is now wrong or obsolete. In parallel, the memory:recall prompt section automatically renders pinned memories first and the most recently updated next, within a character budget-directly in front of the model so it does not have to remember to search. Search then covers anything older than the budget allows.
It is aimed at local-first, offline-minded users who want persistent context across sessions without deploying a vector database or requesting a third-party API key. One SQLite file is the entire cost.
Use Cases
- Retain user preferences and project context across sessions without re-explaining each time
- Pin critical facts into the prompt so the model always sees them
- Provide lightweight long-term memory in fully offline environments
Best For
- Developers who prefer local-first setups and reject external API dependencies
- Users needing cross-session persistent context without deploying a vector database
- Those who want to add memory to Harness at the minimal cost of a single .db file
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.