에이전트 강화
게시자가 제공하는 설치 방식입니다. PlugMCP 엔드포인트가 아닙니다.
원격 URL인가요, 내 컴퓨터에서 실행하는 명령인가요? 두 설치 방식의 차이
PlugMCP에서 실행
A new server you create. It is not this listing's official endpoint. Use the publisher install when you want their URL and their credentials.
실행게시자 제공
This project is a local MCP (Model Context Protocol) server that exposes a small set of tools:
memory.search – semantic search over stored memoriesmemory.save – store a new memorymemory.supersede – mark an old memory as supersededmemory.delete – permanently remove a memory by idLocal Memory MCP Server for Coding Agents
memory.ping – sanity check / version outputWORKSPACE_KEY while keeping one shared local database.memory.search to fetch relevant context.memory.save.memory.supersede or removed via memory.delete.It uses:
@zvec/zvec) as embedded in-process vector database (docs: https://zvec.org/en/docs/)/api/embed with embeddinggemma for embeddings (docs: https://docs.ollama.com/capabilities/embeddings, model: https://ollama.com/library/embeddinggemma)Pull the embedding model:
ollama pull embeddinggemma
bun install
bun run start
This runs an MCP server over stdio.
bun run test
Current tests include:
tests/embed.test.ts – validates Ollama embedding response parsing and error handlingtests/memory-db.test.ts – validates save, search, supersede, and delete on the Zvec-backed storeMEMORY_DB_PATH (default ./data/memory.zvec)OLLAMA_BASE_URL (default http://localhost:11434)OLLAMA_EMBED_MODEL (default embeddinggemma)EMBEDDING_DIM (default 768, must match your embedding model)WORKSPACE_KEY (default default)This repo includes .vscode/mcp.json that registers this server:
bunrun startYou can adjust environment variables in that file.
If you want this MCP server available in all workspaces, add it to your User MCP configuration instead of only .vscode/mcp.json:
MCP: Open User ConfigurationExample (Linux):
{
"servers": {
"local-memory-mcp": {
"type": "stdio",
"command": "bun",
"args": ["--cwd", "/path/to/local-memory-mcp", "run", "start"],
"env": {
"MEMORY_DB_PATH": "/path/to/local-memory-mcp/data/memory.zvec",
"OLLAMA_BASE_URL": "http://localhost:11434",
"OLLAMA_EMBED_MODEL": "embeddinggemma",
"EMBEDDING_DIM": "768",
"WORKSPACE_KEY": "${workspaceFolderBasename}"
}
}
}
}
Notes:
MEMORY_DB_PATH so all projects use the same database.WORKSPACE_KEY=${workspaceFolderBasename} keeps memories separated per project automatically.chat.mcp.autoStart (Experimental) to auto-start/restart MCP servers when needed.Docs:
Add this server to Claude Code as a local stdio MCP server.
This repository already includes:
.mcp.json for project-scoped Claude MCP configurationCLAUDE.md for memory-first agent behavior guidelinesclaude mcp add --transport stdio --scope user \
--env MEMORY_DB_PATH=/absolute/path/to/local-memory-mcp/data/memory.zvec \
--env OLLAMA_BASE_URL=http://localhost:11434 \
--env OLLAMA_EMBED_MODEL=embeddinggemma \
--env EMBEDDING_DIM=768 \
--env WORKSPACE_KEY=default \
local-memory-mcp -- bun --cwd /absolute/path/to/local-memory-mcp run start
claude mcp add --transport stdio --scope project \
--env MEMORY_DB_PATH=./data/memory.zvec \
--env OLLAMA_BASE_URL=http://localhost:11434 \
--env OLLAMA_EMBED_MODEL=embeddinggemma \
--env EMBEDDING_DIM=768 \
--env WORKSPACE_KEY=${PWD##*/} \
local-memory-mcp -- bun run start
Project .mcp.json example:
{
"mcpServers": {
"local-memory-mcp": {
"type": "stdio",
"command": "bun",
"args": ["run", "start"],
"env": {
"MEMORY_DB_PATH": "./data/memory.zvec",
"OLLAMA_BASE_URL": "http://localhost:11434",
"OLLAMA_EMBED_MODEL": "embeddinggemma",
"EMBEDDING_DIM": "768",
"WORKSPACE_KEY": "${PWD##*/}"
}
}
}
}
Notes:
--scope project writes to .mcp.json in the project root.--scope user stores the server in your user Claude configuration.-- before the server command.Useful commands:
claude mcp list
claude mcp get local-memory-mcp
claude mcp remove local-memory-mcp
Docs:
{
"tool": "memory.search",
"arguments": {
"query": "What is our policy for multi-session memory?",
"topK": 8,
"workspaceKey": "my-repo"
}
}
{
"tool": "memory.save",
"arguments": {
"workspaceKey": "my-repo",
"type": "decision",
"summary": "We use zvec with Ollama embeddinggemma for long-term memory.",
"text": "Decision: The Copilot/agent memory sidecar stores vectors in zvec and generates embeddings via Ollama /api/embed using embeddinggemma.",
"tags": ["memory", "zvec", "ollama", "embeddinggemma"],
"importance": 0.8
}
}
{
"tool": "memory.delete",
"arguments": {
"workspaceKey": "my-repo",
"id": 42
}
}
embeddingworkspaceKey, type, summary, etc.)querySync with metadata filters.Run all tests:
bun run test
Current test coverage:
tests/embed.test.ts
/api/embed responses into Float32Arraytests/memory-db.test.ts
save + search behavior with workspace/type filteringsupersede behavior (superseded items are excluded from search)delete behavior and returned payload semanticsLocal Memory MCP Server for Coding Agents
도구
아직 도구 목록이 없습니다. 원격 서버는 첫 상태 확인 후에 도구 목록이 생깁니다.
FAQ
이 페이지에서 Cursor, Claude Desktop, Claude Code 또는 VS Code용 stdio 설정을 복사합니다. PlugMCP 프록시가 아니라 게시자의 설치 방식을 사용합니다.
이 레코드에는 필수 환경 변수가 등록되어 있지 않습니다.
Local Memory은(는) stdio를 통해 로컬에서 실행됩니다.
마지막 확인: 2026-10-28. Local Memory MCP Server for Coding Agents
