Memory & Context Plugins Memory & Context
75 plugins
Plugins that give your agent long-term memory, knowledge persistence, and context management. From simple key-value stores to full knowledge graphs with vector search.
75 plugins
Sort: Featured
-
Listed
Bundle verified
Model-authored context pruning for DeepSeek Harness through the official compaction API.
dsh plugin --profile default add github:aerince/dsh-active-context-pruning -
Listed
Bundle verified
Long-term memory injected at session start: preferences/facts/summaries/knowledge in global and per-project scopes, budgeted recall via `agent/pre-step`, durable JSON store.
dsh plugin --profile default add github:yangyongzhen/dsh-memory -
Listed
Bundle verified
Hana-style assistant manager for DSH: visual preset management (cards, yuan personalities, avatar cropping) plus a full Hana memory stack (rolling summaries, layered compilation, SQLite fact store, deep memory).
dsh plugin --profile default add github:moononnn/DeepSeek-Harness-Hanako-Memory -
Listed
Bundle verified
Turns local markdown and text documents into a queryable knowledge base with hybrid semantic and keyword search, citation verification, and source injection.
dsh plugin --profile default add github:PerryLink/dsh-library -
Listed
Bundle verified
Brings your existing Claude / ChatGPT history into DSH: instant zero-token import of the memories.json inside a Claude export, local Claude Code transcripts with no export needed, and conversations distilled through the model you already configured in DSH. Every memory carries verbatim evidence checked against the source by code — a paraphrase is dropped and counted, and evidence found only in the assistant turns is always routed to review rather than stored.
dsh plugin --profile default add github:Shiye-10Pages/dsh-memory-porter -
Listed
Bundle verified
PostgreSQL-backed durable memory for DSH agents: MCP memory tools plus a Web panel for reviewing memories and candidates.
dsh plugin --profile default add github:AndyYang12345/dsh-butler-memory
Frequently Asked QuestionsFAQ
Memory plugins store knowledge that persists across multiple sessions — facts, entities, and conversation summaries. Session plugins manage the current conversation's state: message history, branching, and import/export.
Most memory plugins add 50-200ms per turn for retrieval. Vector-search plugins are slightly slower but much better at finding relevant context. Key-value stores are fastest.
Yes. A common pattern is context-compression for the current session plus a vector-recall plugin for long-term knowledge.