AI Skill Hub 推荐使用:Flex 是一款优质的MCP工具。AI 综合评分 7.5 分,在同类工具中表现稳健。如果你正在寻找可靠的MCP工具解决方案,这是一个值得深入了解的选择。
Flex 是一款遵循 MCP(Model Context Protocol)标准协议的 AI 工具扩展。通过 MCP 协议,它可以让 Claude、Cursor 等主流 AI 客户端直接访问和操作外部工具、数据源和服务,实现 AI 能力的无缝扩展。无论是文件操作、数据库查询还是 API 调用,都可以通过自然语言在 AI 对话中直接触发,极大提升生产效率。
Flex 是一款遵循 MCP(Model Context Protocol)标准协议的 AI 工具扩展。通过 MCP 协议,它可以让 Claude、Cursor 等主流 AI 客户端直接访问和操作外部工具、数据源和服务,实现 AI 能力的无缝扩展。无论是文件操作、数据库查询还是 API 调用,都可以通过自然语言在 AI 对话中直接触发,极大提升生产效率。
# 方式一:通过 Claude Code CLI 一键安装
claude skill install https://github.com/damiandelmas/flex
# 方式二:手动配置 claude_desktop_config.json
{
"mcpServers": {
"flex": {
"command": "npx",
"args": ["-y", "flex"]
}
}
}
# 配置文件位置
# macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
# Windows: %APPDATA%/Claude/claude_desktop_config.json
# 安装后在 Claude 对话中直接使用 # 示例: 用户: 请帮我用 Flex 执行以下任务... Claude: [自动调用 Flex MCP 工具处理请求] # 查看可用工具列表 # 在 Claude 中输入:"列出所有可用的 MCP 工具"
// claude_desktop_config.json 配置示例
{
"mcpServers": {
"flex": {
"command": "npx",
"args": ["-y", "flex"],
"env": {
// "API_KEY": "your-api-key-here"
}
}
}
}
// 保存后重启 Claude Desktop 生效
<p align="center"> <img src="assets/banner.png" alt="flex" width="100%"> </p>
**most memory systems start working after you install them.** flex works retroactively — the moment you install, your existing sessions are queryable. Ask how you set up the Cloudflare tunnel yesterday, why a release script changed, or what session created a file.
sessions aren't plain documents. They have prompts, replies, tool calls, file edits, repos, projects, and sub-agents. flex keeps that structure, so your agent filters before semantic scoring instead of asking vector search to guess what matters. Every answer stays attached to its source evidence — the session it came from, the files and repos it touched, and where to go next for the full trace.
vector search usually returns similar content and stops. flex lets your agent compose SQL, semantic, and keyword search with operators for suppression, diversity, recency, and trajectory — architecture work but not changelogs, recent auth work but not oauth docs, a diverse sample instead of ten near-duplicates.
- MCP server: one read-only query tool, flex_search — the primary interface for agents. - local SQLite databases: source-specific tables, views, saved queries, and runtime docs your agent can inspect. - CLI: initialize sources and inspect health. - flexvec: SQL vector retrieval kernel with suppression, diversification, decay, and trajectory operators. - worker: background refresh for local coding-agent memory.
Claude Code:
curl -sSL https://getflex.dev/install.sh | bash -s -- claude-code
Codex CLI:
curl -sSL https://getflex.dev/install.sh | bash -s -- codex
Obsidian / Markdown:
curl -sSL https://getflex.dev/install.sh | bash -s -- obsidian
MCP is transport, not topology. The agent sees a single read-only tool — flex_search against a named cell — and retrieval happens inside the cell (see how retrieval works). The flex core CLI is for installation and operations: initialize sources, rebuild cells, inspect health. The durable artifact is the cell itself — one local SQLite file under ~/.flex/cells/, with no hosted service in the query path.
Extension modules enrich any cell with shared structure — they don't add a source or a tool.
A module installs by creating tables with the convention prefixes and uninstalls by dropping them — no registration, no coupling. A cell without a given module still has full retrieval; those columns are simply absent. SOMA and the graph enrichments are modules in exactly this sense — ordinary SQLite tables and columns, queried as SQL, never a separate tool.
Flex是一个有潜力的开源MCP工具,提供AI智能搜索和检索功能
AI Skill Hub 为第三方内容聚合平台,本页面信息基于公开数据整理,不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后,再部署至生产环境,并做好必要的安全评估。
✅ MIT 协议 — 最宽松的开源协议之一,可自由商用、修改、分发,仅需保留版权声明。
总体来看,Flex 是一款质量良好的MCP工具,在同类工具中具备一定竞争力。AI Skill Hub 将持续追踪其更新动态,建议收藏备用,结合自身场景选择合适时机引入使用。
| 原始名称 | flex |
| 原始描述 | 开源MCP工具:Search and retrieval for AI agents。⭐9 · Python |
| Topics | mcpai-agentsembeddingsknowledge-base |
| GitHub | https://github.com/damiandelmas/flex |
| License | MIT |
| 语言 | Python |
收录时间:2026-05-30 · 更新时间:2026-05-30 · License:MIT · AI Skill Hub 不对第三方内容的准确性作法律背书。
选择 Agent 类型,复制安装指令后粘贴到对应客户端