Skip to main content
Exa helps you search the web intelligently, pull high quality content, discover related pages, and get citation-backed answers. Use it to find trusted sources faster, gather page contents and metadata at scale, and drive research tasks with clear citations.
Leverage neural web search, instant content extraction, similar-page discovery, and LLM answers with citations for fast, reliable research.

​
How to Use MCP Nodes

​
What is Exa MCP?

The Exa MCP creates a customized node that understands Exa.ai and responds to natural language prompts. Describe what you need, and the node executes Exa tools to search, fetch contents, find similar pages, answer questions, or run research tasks. Results come back as structured data that you can pass to the next step in your workflow.

​
What Can It Do for You?

  • Find high quality, relevant web results for any topic using neural and keyword search
  • Fetch full page contents, summaries, and metadata for URLs at scale
  • Discover similar pages to a source URL to broaden research coverage
  • Get citation-backed answers and run asynchronous research tasks with clear statuses and outputs

​
Available Tools

ToolWhat It DoesExample Use
SearchIntelligently search the web using Exa’s neural and keyword search. Returns relevant results with optional content extraction including text, highlights, and summaries.”Search the web for search term and return structured data with title, url, and a short snippet for the top result count results”
Get ContentsGet full page contents, summaries, and metadata for a list of URLs. Returns instant results from cache with automatic live crawling as fallback.”For each url in urls, fetch page content and return structured data with title, url, author, published date, summary, and main text”
Find SimilarFind similar links to a provided URL using Exa’s neural similarity search to discover related content.”Given source url, find similar pages and return structured data with title and url for the top result count matches”
AnswerGet an LLM-generated answer informed by Exa search results. Returns a direct answer for specific queries or a detailed summary with citations for open-ended queries.”Answer the question question using reliable web sources and return structured data with answer text and citations including title and url”
Create Research TaskStart an asynchronous research task that explores the web, gathers sources, synthesizes findings, and returns results with citations. Returns immediately with a researchId for polling.”Create a research task for topic and return structured data with researchId and initial status”
Get Research TaskRetrieve the status and results of a previously created research task. Poll until status is completed, failed, or canceled.”Using researchId, get the research task status and if completed return structured data with key findings and citations including title and url”

​
Credit Costs

ToolCredits Per Use
Get Contents3 per item
Find Similar5 per item
Answer10 credits
Create Research Task5 credits
Get Research Task5 credits

​
How to Use

1

Create Your Exa MCP Node

Go to your node library, search for Exa, and click “Create a node with AI”
2

Add Your Prompt

Drag the Exa MCP node to your canvas and add your prompt in the text box.
3

Test Your Node

Run the node to see the results. If it works as expected, you’re all set. If you need adjustments, check the troubleshooting tips below.
4

Save and Reuse

Once your Exa MCP node is working, save it to your library. You can now use this customized node in any workflow.

​
Example Prompts

Here are some prompts that work well with Exa MCP: Targeted Web Search: 
Search the web for `search term` and return structured data with title, url, and a short snippet for the top `result count` results
Fetch Page Contents and Metadata: 
For each url in `urls`, get the page contents and return structured data with title, url, author, published date, summary, and main text
Discover Related Sources: 
Find similar pages to `source url` and return structured data with title and url for the top `result count` matches
Citation-backed Answer: 
Answer the question `question` and return structured data with answer text and citations including title and url
Asynchronous Research - Create: 
Create a research task for `topic` and return structured data with researchId and initial status
Start simple and focused. For broad questions, begin with Search to shortlist sources, then use Get Contents on the top URLs, and finally ask Answer for a concise, citation-backed synthesis. Saving each step as its own node keeps your workflows fast and reusable.

​
Troubleshooting

If your Exa MCP node isn’t working as expected, try these best practices:

​
Keep Prompts Simple and Specific

  • Good: “Search the web for search term and return title and url”
  • Less Efficient: “Search the web for topic, fetch the contents of each result, summarize them, and write a recommendation”
While the longer prompt might run, it’s more efficient to break it into separate nodes. Exa MCP works best with focused, single-action prompts.

​
Match What Exa Can Do

  • Good: “Given source url, find similar pages and return title and url”
  • Less Efficient: “Find similar pages to source url and add them to a Google Sheet”
Exa MCP excels at web search, content retrieval, similar-page discovery, answers, and research tasks. For writing to spreadsheets, combine it with Google Sheets nodes in your workflow.

​
Break Complex Tasks Into Steps

Trying to do everything in one prompt can sometimes lead to timeouts or unexpected results: 
Search for all articles about `topic`, get contents for each url, synthesize findings, and draft an email using `email template`
A more efficient approach is to split this into smaller, focused nodes:
1

Step 1: Search the Web

Search the web for topic and return structured data with title, url, and snippet for the top result count results
2

Step 2: Get Page Contents

For each url in urls, fetch contents and return structured data with title, url, summary, and main text
3

Step 3: Generate Summary or Email

Using the page summaries, create a concise overview with Ask AI and return structured data with key points or an email draft
In your workflow, connect these nodes sequentially. The output from Step 1 becomes the input for Step 2, and Step 2’s summaries feed into Step 3.

​
Focus on Data Retrieval

Exa MCP is great at getting information from the web. For analysis or content creation, connect it to other nodes. Example:
  • Good prompt: “For each url in urls, return structured data with title, published date, and summary”
  • Less Efficient: “For each url in urls, summarize the article and write a LinkedIn post”
Use Ask AI for writing posts or deeper analysis, and Categorizer for classification. Keep Exa focused on finding and retrieving data, then hand off to the right node.

​
Troubleshooting Node Creation

If you’re seeing empty outputs in the node creation window (or if you’ve already created the node, hover over it and click “Edit”), use the chat interface to prompt the AI to add debug logs and verify the API response. Specifically mention that you received empty outputs.
In the node creation window (or if you’ve already created the node, hover over it and click “Edit”), use the chat interface to describe in detail what you expected versus what you received.
First click “Fix with Gummie”. If multiple attempts fail, simplify your prompt or contact support.
MCP node creation often requires a few tweaks. Use the chat interface in the node creation window to refine filters, output fields, or pagination. The AI will adjust the node based on your feedback.

​
Need More Help?

⌘I