LlamaCloud MCP Server

@run-llama

代码仓库

基本信息

分类

AI 与智能体

许可证

MIT

传输方式

stdio

作者

run-llama

关于 LlamaCloud MCP Server

A MCP server connecting to managed indexes on LlamaCloud

#llamacloud#mcp-server#managed-index

LlamaCloud MCP Server 配置

使用下面的配置,将此服务器添加到你的 MCP 客户端。

{
  "mcpServers": {
    "llamacloud": {
      "command": "npx",
      "args": [
        "-y",
        "@llamaindex/mcp-server-llamacloud",
        "--index",
        "10k-SEC-Tesla",
        "--description",
        "10k SEC documents from 2023 for Tesla",
        "--topK",
        "5",
        "--index",
        "10k-SEC-Apple",
        "--description",
        "10k SEC documents from 2023 for Apple"
      ],
      "env": {
        "LLAMA_CLOUD_API_KEY": "<YOUR_API_KEY>"
      }
    }
  }
}

LlamaCloud MCP Server 工具

generates tool names like `get_information_index_name` based on index names

概览

What is LlamaCloud MCP Server?

LlamaCloud MCP Server is a TypeScript-based MCP server that connects to multiple managed indexes on LlamaCloud. It creates separate MCP tools for each index, allowing users to query their LlamaCloud documents through an MCP client.

How to use LlamaCloud MCP Server?

Configure the server in your MCP client (e.g., Claude Desktop, Windsurf, or Cursor) by adding an entry with command npx -y @llamaindex/mcp-server-llamacloud and passing --index, --description, and optionally --topK arguments for each index. Set the LLAMA_CLOUD_API_KEY environment variable to your API key.

Key features of LlamaCloud MCP Server

Creates a separate tool for each defined index
Each tool provides a query parameter for searching
Auto-generates tool names like get_information_index_name
Supports optional --topK to limit results per tool

Use cases of LlamaCloud MCP Server

Querying 10k SEC documents for multiple companies from a single server
Searching indexed knowledge bases through an MCP client interface
Managing and querying multiple LlamaCloud projects within one agent

常见问题

How do I configure the server for my MCP client?

Add a JSON entry to your MCP client config with command `npx -y @llamaindex/mcp-server-llamacloud`, provide `--index` and `--description` pairs, and set `LLAMA_CLOUD_API_KEY` in the `env` block.

Where can I find my MCP client configuration file?

On macOS: `~/Library/Application Support/Claude/claude_desktop_config.json`. On Windows: `%APPDATA%/Claude/claude_desktop_config.json`.

How do I define multiple indexes (tools)?

Repeat the `--index` and `--description` arguments in the `args` array. Each pair creates a new tool. Optionally add `--topK` after a pair to limit its results.

What environment variables are required?

`LLAMA_CLOUD_API_KEY` is required. `LLAMA_CLOUD_PROJECT_NAME` is optional and defaults to `Default` if not set.

How can I debug the server?

Use the MCP Inspector by running `npm run inspector` from the project directory. It provides a URL with debugging tools for stdio-based MCP servers.

AI 与智能体分类下的更多 MCP 服务器

Ranla

Kurtis Tryber

AI growth agent for software businesses. Ranla ships pages that rank on Google, gets your product named in AI answers, replies on Reddit and sends lifecycle email. Ask how growth is going, have Ranla draft the work and a

Voibe Transcription

Ayush Chaturvedi

Give ears to Claude/Openclaw/Hermes/Codex/Grok Bot. Voibe turns recordings into text your AI agent can work with. Ask your agent to transcribe a meeting, interview, call, lecture, podcast episode or voice memo. The raw

Krafters

Krafters (Romain Duclos)

Krafters is a French club of specialized AIs for freelancers and small businesses. Its connector opens the member's AI kits for Facebook and Instagram ads, SEO/GEO and short videos directly in Claude, and gives the next

Simplepages

Build and edit landing pages and websites from your AI assistant, and pull how they are performing.

AgentRisk M2M

Neurobyteio

AgentRisk M2M 🛡️ The Security & Trust Layer for Autonomous Agents on Base. Stop letting your trading agents get rekt by stealth honeypots, dynamic sell-blocks, and rugpulls that pass every static check. ⚡ The Problem