mcp-jina mcp

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

Jina AI services: web reader (r.jina.ai), search (s.jina.ai), embeddings, reranking and classification. API key optional for free tier reader/search.

jinareaderembeddingsrerankerrag

Input Parameters

Parameter	Type	Description
operationrequired	string	read: extract clean text from URL; search: web search; embed: compute embeddings; rerank: rerank documents for a query; classify: classify texts. One of: `read`, `search`, `embed`, `rerank`, `classify`.
apiKeyoptional	string	Jina API key (optional for read/search free tier).
urloptional	string	URL to read. Required for read.
queryoptional	string	Search query or reranking query.
textsoptional	string	JSON array of text strings for embed, rerank or classify operations.
modeloptional	string	Model name for embed/rerank/classify (e.g. jina-embeddings-v3).
documentsoptional	string	JSON array of documents to rerank against the query.
topNoptional	integer	Top N results to return for rerank. Default: `5`.
labelsoptional	string	Comma-separated class labels for classify.
returnFulloptional	boolean	Return full text content for read (default: false — returns summary). Default: `False`.
withLinksoptional	boolean	Include extracted links in read output. Default: `False`.
withImagesoptional	boolean	Include image URLs in read output. Default: `False`.
timeoutoptional	integer	Request timeout in seconds (default: 30). Default: `30`.

Output Fields

Field	Type	Description
content	string
results	array[object]
embeddings	array[array]
reranked	array[object]
classes	array[object]
error	string

Examples

Read a webpage

// Input
{
  "operation": "read",
  "url": "https://github.com/anthropics/anthropic-sdk-python"
}

// Output
{
  "content": "# anthropic-sdk-python\n..."
}

Install & Discovery

Install

ppm install mcp-jina

Get JSON Schema

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

Discover by keyword

GET /v1/mcp/discover?q=jina

Discovery hint: Install with ppm install mcp-jina or invoke remotely via POST /v1/invoke/mcp-jina on the MCP Service.

PascalAI Usage

uses toolslib;
var Tool := LoadTool('mcp-jina');
var R := Tool.Call(JsonObj(['operation','read','url','https://example.com']));
Writeln(R.ToJSON);