Skip to main content
Glama

Letta MCP 服务器

npm version Rust Tests Coverage Docker Build License: MIT

一个高性能的 Model Context Protocol (MCP) 服务器,专为 Letta AI 构建,使用 Rust 和 TurboMCP 框架开发。

特性

  • 7 个整合工具:使用判别器模式涵盖 103 种操作

  • 高性能 -- Rust 编写,内存占用约 10-30MB,启动时间 <500ms,通过 tokio::join!buffer_unordered 实现并行扇出

  • 双重传输 -- 支持 stdio(Claude Desktop、Cursor 等)和 HTTP(生产环境部署)

  • 响应大小优化 -- 响应大小减少 68-96%,提升 LLM 上下文效率

  • 类型收窄模式 -- 具有特定 object/array 类型及参数描述(无通用 Value 参数)

  • 多平台支持 -- macOS、Linux、Windows (x64 和 arm64)

  • 兼容 Letta 0.16.x -- 支持存档、对话、MCP 服务器 v2

  • 符合 MCP 2025-11-25 标准 -- 支持带 SSE 的流式 HTTP 传输

  • 编译时路由 -- 通过 TurboMCP v3 宏实现零运行时开销的工具分发

Related MCP server: A2A Client MCP Server

快速入门

npm (推荐)

npm install -g letta-mcp-server

系统会自动安装适合您平台的二进制文件。

平台

软件包

macOS Intel

letta-mcp-darwin-x64

macOS Apple Silicon

letta-mcp-darwin-arm64

Linux x64

letta-mcp-linux-x64

Linux arm64

letta-mcp-linux-arm64

Windows x64

letta-mcp-windows-x64

Docker

docker pull ghcr.io/oculairmedia/letta-mcp-server-rust:rust-latest

docker run -d \
  -p 6507:6507 \
  -e LETTA_BASE_URL=http://your-letta-instance:8283 \
  -e LETTA_PASSWORD=your-password \
  -e TRANSPORT=http \
  --name letta-mcp \
  ghcr.io/oculairmedia/letta-mcp-server-rust:rust-latest

Docker Compose

services:
  letta-mcp:
    image: ghcr.io/oculairmedia/letta-mcp-server-rust:rust-latest
    container_name: letta-mcp
    restart: unless-stopped
    ports:
      - '6507:6507'
    environment:
      LETTA_BASE_URL: ${LETTA_BASE_URL}
      LETTA_PASSWORD: ${LETTA_PASSWORD}
      TRANSPORT: http
      PORT: 6507
      RUST_LOG: info
    env_file:
      - .env
    healthcheck:
      test: ['CMD-SHELL', "timeout 1 bash -c '</dev/tcp/localhost/6507' || exit 1"]
      interval: 30s
      timeout: 10s
      retries: 3

环境变量

变量

必需

默认值

描述

LETTA_BASE_URL

Letta API URL (例如 http://localhost:8283)

LETTA_PASSWORD

Letta API 密码

TRANSPORT

stdio

传输模式: stdiohttp

PORT

6507

HTTP 端口 (当 TRANSPORT=http 时)

RUST_LOG

info

日志级别: debug, info, warn, error

RUST_BACKTRACE

0

启用回溯 (01)

LETTA_MCP_MAX_VALUE_LEN

500

块/段落/搜索值的最大字符数

LETTA_MCP_CORE_MEMORY_PREVIEW_LEN

200

核心内存预览的最大字符数

可用工具

该服务器提供 7 个整合工具,包含 103 种操作

工具

操作数

描述

letta_agent_advanced

28

智能体生命周期、消息传递、对话、上下文、导出/导入

letta_memory_unified

24

核心内存、块、存档段落、存档、搜索

letta_tool_manager

13

工具 CRUD、挂载/卸载、批量操作

letta_source_manager

15

数据源、文件、段落、附件

letta_job_monitor

4

任务跟踪、取消、主动监控

letta_file_folder_ops

8

文件会话、文件夹管理

letta_mcp_ops

11

MCP 服务器生命周期、工具发现、v2 API 支持

工具操作

letta_agent_advanced (28 种操作)

list, create, get, update, delete, search, list_tools, send_message,
export, import, clone, get_config, bulk_delete, context, reset_messages,
summarize, stream, async_message, cancel_message, preview_payload,
search_messages, get_message, count, list_conversations,
get_conversation, send_conversation_message, cancel_conversation,
compact_conversation

letta_memory_unified (24 种操作)

get_core_memory, update_core_memory, get_block_by_label, list_blocks,
create_block, get_block, update_block, attach_block, detach_block,
list_agents_using_block, search_archival, list_passages, create_passage,
update_passage, delete_passage, search_memory, list_archives,
get_archive, create_archive, update_archive, delete_archive,
attach_archive, detach_archive, list_agents_using_archive

letta_tool_manager (13 种操作)

list, get, create, update, delete, upsert, attach, detach, bulk_attach,
generate_from_prompt, generate_schema, run_from_source, add_base_tools

letta_source_manager (15 种操作)

list, get, create, update, delete, count, attach, detach, list_attached,
upload, delete_files, list_files, list_folders, get_folder_contents,
list_agents_using

letta_job_monitor (4 种操作)

list, get, cancel, list_active

letta_file_folder_ops (8 种操作)

list_files, open_file, close_file, close_all_files, list_folders,
attach_folder, detach_folder, list_agents_in_folder

letta_mcp_ops (11 种操作)

add, update, delete, test, connect, resync, list_servers, list_tools,
register_tool, execute, attach_mcp_server

响应大小优化

Rust 实现包含针对 LLM 上下文效率的显著响应大小优化:

操作

优化

大小缩减

智能体列表

默认分页 (15 项),摘要模式

68-85%

工具列表

默认分页 (25 项),截断描述

70-90%

内存块

列表模式下排除繁重字段

60-80%

源列表

摘要模式,分页

75-95%

分页

所有列表操作均支持分页:

{
  "operation": "list",
  "pagination": {
    "limit": 25,
    "offset": 0
  }
}

摘要模式 vs 完整模式

列表操作默认返回摘要数据。使用带有特定 ID 的 get 操作获取完整详情:

// Summary (default for list)
{
  "id": "agent-123",
  "name": "My Agent",
  "model": "gpt-4",
  "tool_count": 5
}

// Full (with get operation)
{
  "id": "agent-123",
  "name": "My Agent",
  "model": "gpt-4",
  "system_prompt": "...",
  "tools": [...],
  "memory_blocks": [...]
}

MCP 客户端配置

Claude Desktop

{
  "mcpServers": {
    "letta": {
      "command": "letta-mcp",
      "env": {
        "LETTA_BASE_URL": "http://localhost:8283",
        "LETTA_PASSWORD": "your-password"
      }
    }
  }
}

Cursor / Windsurf

{
  "mcpServers": {
    "letta": {
      "command": "letta-mcp",
      "env": {
        "LETTA_BASE_URL": "http://localhost:8283",
        "LETTA_PASSWORD": "your-password"
      }
    }
  }
}

OpenCode (HTTP)

{
  "mcp": {
    "letta-mcp": {
      "type": "remote",
      "url": "http://localhost:6507/mcp",
      "enabled": true
    }
  }
}

从源码构建

前置要求

  • Rust nightly (支持 2024 版本) - 通过 rustup 安装

  • Docker (可选,用于容器化构建)

本地构建

git clone https://github.com/oculairmedia/Letta-MCP-server.git
cd Letta-MCP-server

cargo build --release

LETTA_BASE_URL=http://your-letta:8283 \
LETTA_PASSWORD=your-password \
./target/release/letta-server

Docker 构建

docker build -f Dockerfile.rust -t letta-mcp .

docker run -d \
  -p 6507:6507 \
  -e LETTA_BASE_URL=http://your-letta:8283 \
  -e LETTA_PASSWORD=your-password \
  -e TRANSPORT=http \
  letta-mcp

架构

letta-server/
├── src/
│   ├── main.rs                        # Entry point, transport selection
│   ├── lib.rs                         # Server init, compile-time tool registration
│   └── tools/
│       ├── mod.rs                     # Tool module exports
│       ├── response_utils.rs          # Unified ToolResponse, pagination helpers
│       ├── validation_utils.rs        # require_field, require_id, sdk_err
│       ├── id_utils.rs               # ID parsing utilities
│       ├── agent_advanced/            # Agent operations (28 ops)
│       │   ├── crud.rs               #   list, create, get, update, delete, search
│       │   ├── messaging.rs           #   send_message, stream, async, preview
│       │   ├── conversations.rs       #   list, get, send, cancel, compact
│       │   └── management.rs          #   export, import, clone, config, bulk_delete
│       ├── memory_unified/            # Memory operations (24 ops)
│       │   ├── core.rs               #   get/update core memory
│       │   ├── blocks.rs             #   block CRUD, attach/detach
│       │   ├── passages.rs           #   archival passage CRUD
│       │   ├── archives.rs           #   archive CRUD, attach/detach
│       │   └── search.rs             #   search_archival, search_memory
│       ├── tool_manager.rs            # Tool operations (13 ops)
│       ├── source_manager.rs          # Source operations (15 ops)
│       ├── job_monitor.rs             # Job operations (4 ops)
│       ├── file_folder_ops.rs         # File/folder operations (8 ops)
│       └── mcp_ops.rs                # MCP server v2 operations (11 ops)
├── tests/                             # Integration tests
└── Cargo.toml

关键依赖

  • TurboMCP v3 -- 具有编译时路由和流式 HTTP 的 MCP 框架

  • letta-rs -- 供应商提供的 Rust Letta API 客户端 (0.16.x)

  • Tokio -- 异步运行时

  • Futures -- 流组合器 (buffer_unordered 用于并行扇出)

  • Serde -- 序列化/反序列化

故障排除

连接被拒绝 (Connection Refused)

  1. 确保服务器正在运行: docker ps | grep letta-mcp

  2. 检查日志: docker logs letta-mcp

  3. 验证端口是否可访问: curl http://localhost:6507/mcp

认证错误

  1. 验证 LETTA_BASE_URL 是否指向您的 Letta 实例

  2. 检查 LETTA_PASSWORD 是否正确

  3. 确保从容器中可以访问 Letta 服务器

工具未找到

  1. 通过 MCP 列出可用工具: tools/list

  2. 验证您是否使用了正确的操作名称 (例如 list 而不是 list_agents)

日志

# View server logs
docker logs -f letta-mcp

# Enable debug logging
RUST_LOG=debug letta-mcp

贡献

  1. Fork 本仓库

  2. 创建功能分支: git checkout -b feature/my-feature

  3. 进行更改并添加测试

  4. 运行测试: cargo test

  5. 提交 Pull Request

许可证

MIT 许可证 - 详情请参阅 LICENSE 文件。

相关项目

Install Server
A
license - permissive license
A
quality
A
maintenance

Maintenance

Maintainers
4dResponse time
2wRelease cycle
8Releases (12mo)
Commit activity
Issues opened vs closed

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Servers

View all related MCP servers

Related MCP Connectors

  • MCP server exposing the Backtest360 engine API as tools for AI agents.

  • MCP server for AI dialogue using various LLM models via AceDataCloud

  • Cloud-hosted MCP server for durable AI memory

View all MCP Connectors

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/oculairmedia/Letta-MCP-server'

If you have feedback or need assistance with the MCP directory API, please join our Discord server