Jinaai Grounding

Provides real-time web grounding for LLMs via Jina.ai Grounding API, integrated into MCP for structured, factual responses.

javascript

1

GitHub Stars

javascript

Language

6 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "spences10-mcp-jinaai-grounding": {
      "command": "node",
      "args": [
        "-y",
        "mcp-jinaai-grounding"
      ],
      "env": {
        "JINAAI_API_KEY": "YOUR_API_KEY"
      }
    }
  }
}

You can run a dedicated MCP server that grounds language model outputs to real-time web content using Jina.ai Grounding. This server provides precise, up-to-date web grounding for LLM prompts and responses, helping you verify facts and enrich answers with relevant sources.

How to use

You interact with the grounding server through an MCP client. Start the server locally and reference it by its MCP configuration name. Use the grounding tool to attach real-time web content to your LLM responses, specify how you want the grounding to be formatted, and control behavior like caches, redirects, and image handling. The core tool is designed to ground a given query against web content and return structured, LLМ-friendly results that emphasize relevance and accuracy.

How to install

Prerequisites you need before installation:

# Ensure you have Node.js and a package manager installed
node -v
npm -v
pnpm -v

Install the package manager for this MCP server’s workspace, then install dependencies, build, and run in development mode.

pnpm install
pnpm run build
pnpm run dev

To run using the standard MCP setup with a local execution, you will typically configure the MCP client to invoke the server using a Node runtime and the provided entry path.

Additional configuration and usage notes

Environment variable you must provide for authentication with Jina.ai Grounding is JINAAI_API_KEY. Supply this in your MCP client configuration or environment when starting the server.

The server exposes a tool named ground_content that grounds LLM responses with real-time web content. You can tailor the grounding behavior with parameters such as query, no_cache, format, token_budget, stream, gather_links, gather_images, image_caption, enable_iframe, enable_shadow_dom, and resolve_redirects to suit your needs.

Security and best practices

Keep your JINAAI_API_KEY secret. Do not embed it in client-side code or public repositories. Use environment variables or secret management in your deployment environment. When possible, restrict the API key’s scope and rotate keys periodically.

Troubleshooting

If grounding results are not appearing, verify that the API key is valid and that the MCP client is correctly configured to invoke the local server. Check for network or process startup errors in your development console and ensure that the server process is running with the expected command and arguments.

Examples and tips

Use ground_content to attach real-time web content to your prompts. Adjust options like format to receive json for structured data or text for human-friendly output. Enable gather_links and gather_images if you want supplementary citations or media at the end of the response.

Available tools

ground_content

Ground LLM responses with real-time web content using Jina.ai Grounding. Supports options like cache control, formatting, token budgeting, and optional gathering of links/images.