dsh-vector-memory
Run the following command in DeepSeek Harness:
dsh plugin install Bryan-cmf/dsh-vector-memory
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install Bryan-cmf/dsh-vector-memory in your terminal to install the plugin; full source is available at https://github.com/Bryan-cmf/dsh-vector-memory .
About this plugin
Persisting agent memory across sessions used to mean standing up Qdrant, BGE-m3, and a Docker pipeline—even to store a handful of paragraphs. dsh-vector-memory removes all of that infrastructure: it runs natively inside DeepSeek Harness, writing to the DSH storageDomain JSON backend with zero external services and zero setup scripts. It just works.
The plugin exposes three native tools. mem_save writes a tagged, TTL-aware memory entry and returns its id. mem_search scores hits by deterministic keyword matching plus recency decay, returning a ranked list of content, tags, and timestamp. mem_health reports record count, expired count, storage-domain status, and the active TTL policy. Other plugins can inject the vectorMemory service and call save / search / health directly. On the client side, a live Memory view tab is registered so you can inspect everything saved and retrieved in the current session at a glance.
It is ideal for developers building multi-turn agents on DSH who want a lightweight, persistent memory layer without owning a vector database or an embedding model. Configuration is reduced to two knobs—default TTL of 90 days and 10 results per search—making it one of the lowest-friction ways to add durability to an agent loop. Note: this suite was merged into dsh-insights in August 2026; head to the dsh-insights repository for all ongoing development.
Use Cases
- Agents retain user preferences and context across sessions
- Persist intermediate task state across multi-turn conversations
- Expose a shared vectorMemory service to other plugins
Best For
- Developers building multi-turn agents on DSH
- Teams wanting persistent memory with zero infrastructure overhead
- Service builders needing cross-plugin shared state
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.