mcp-exa mcp

v1.0.0 · MCP Tool · search · registry.pascalai.org

Exa AI-powered semantic web search (exa.ai). Requires API key. Operations: search (semantic/neural search), find_similar (pages similar to URL), get_contents (retrieve full content by URL or IDs).

exasearchneuralsemanticweb

Input Parameters

ParameterTypeDescription
operationrequired string search: semantic web search; find_similar: find pages similar to a URL; get_contents: retrieve full page contents. One of: search, find_similar, get_contents.
apiKeyrequired string Exa API key.
queryoptional string Search query. Required for search.
numResultsoptional integer Number of results (default: 10). Default: 10.
searchTypeoptional string Search type (default: neural). One of: neural, keyword, auto. Default: neural.
categoryoptional string Content category filter (e.g. news, research paper, tweet).
startPublishedDateoptional string ISO 8601 date — only return results published after this date.
endPublishedDateoptional string ISO 8601 date — only return results published before this date.
includeDomainsoptional string Comma-separated list of domains to include (e.g. 'github.com,arxiv.org').
excludeDomainsoptional string Comma-separated list of domains to exclude.
includeTextoptional boolean Include full page text in results. Default: False.
urloptional string URL for find_similar operation.
idsoptional string Comma-separated Exa document IDs for get_contents.
maxCharsoptional integer Max characters of text to return per result. Default: 2000.

Output Fields

FieldTypeDescription
results array[object]
count integer
error string

Examples

Semantic search

// Input
{
  "operation": "search",
  "apiKey": "your-key",
  "query": "latest AI research",
  "numResults": 5
}

// Output
{
  "count": 5,
  "results": [
    {
      "title": "...",
      "url": "...",
      "score": 0.95
    }
  ]
}

Install & Discovery

Install

ppm install mcp-exa

Get JSON Schema

GET /v1/packages/mcp-exa/1.0.0/schema

Discover by keyword

GET /v1/mcp/discover?q=exa
Discovery hint: Install with ppm install mcp-exa or invoke remotely via POST /v1/invoke/mcp-exa on the MCP Service.

PascalAI Usage

uses toolslib;
var Tool := LoadTool('mcp-exa');
var R := Tool.Call(JsonObj(['operation','search','apiKey','your-key','query','AI news']));
Writeln(R.ToJSON);