Jina AI Remote

You can access Jina’s remote MCP server to tap into Reader, Embeddings, and Reranker APIs, plus a suite of web, image, and content tools via MCP. This setup lets you combine powerful AI capabilities with lightweight client configurations, enabling advanced search, content extraction, and question-answering workflows without hosting the full MCP locally.

How to use

Configure your MCP client to connect to the remote MCP server endpoints. Use the HTTP connection for direct remote access or run a local proxy to connect to the remote server. In both cases you’ll gain access to tools for reading web content, searching the web and arXiv, handling images, and performing reranking and deduplication to produce high-quality results.

How to install

Prerequisites: you need Node.js and npm installed on your machine. If you prefer a local proxy, you will also use npx to run the remote MCP proxy.

Step 1. Install Node.js and npm if you don’t have them already. Visit the Node.js website and install the LTS version suitable for your operating system.

Step 2. Prepare your MCP client configuration. Choose HTTP direct access or a local proxy as shown below.

Step 3. Run your chosen configuration snippet in your MCP-enabled client. Use the exact command or JSON block as shown in the examples.

Configuration and usage notes

Two connection methods are provided for the remote MCP server. Use the HTTP method for direct access and a local stdio proxy if your client requires a local process to connect.

Troubleshooting and tips

If you encounter tool calling loops or missing tool definitions, refresh your MCP client’s tool cache or re-add the jina_mcp configuration to your client. For Windows users, ensure arguments containing spaces are passed correctly to the proxy command by avoiding spaces around separators in environment variables where required.

Security and keys

If your client supports API keys, include your Jina API key in the Authorization header as a Bearer token to access protected tools. Treat your API key as a secret and avoid exposing it in shared configurations.

Examples of common tool access

You can perform contextual querying, web and arXiv searches, image searches, and content extraction with markdown-ready results. You can also expand queries, parallelize reads and searches, and deduplicate results both in strings and images to improve relevance and conciseness.

Available tools

primer

Get current contextual information for localized, time-aware responses.

read_url

Extract clean, structured content from web pages as markdown via Reader API.

capture_screenshot_url

Capture high-quality screenshots of web pages via Reader API.

guess_datetime_url

Analyze web pages for last update/publish datetime with confidence scores.

search_web

Search the entire web for current information and news via Reader API.

search_arxiv

Search academic papers and preprints on arXiv repository via Reader API.

search_images

Search for images across the web via Reader API.

expand_query

Expand and rewrite search queries based on the query expansion model via Reader API.

parallel_read_url

Read multiple web pages in parallel for efficient content extraction via Reader API.

parallel_search_web

Run multiple web searches in parallel for topic coverage via Reader API.

parallel_search_arxiv

Run multiple arXiv searches in parallel for comprehensive academic coverage.

sort_by_relevance

Rerank documents by relevance to a query via Reranker API.

deduplicate_strings

Get top-k semantically unique strings via Embeddings API and submodular optimization.

deduplicate_images

Get top-k semantically unique images via Embeddings API and submodular optimization.