Spark
Sign In
MCP Connector Featured

Connect to Spark AI Catalog

Spark MCP Server lets AI assistants search and retrieve AI assets from the Entire VC Spark catalog, hosted or run locally.

Connect
Pavel Rogozhin
Pavel Rogozhin
97
Spark score
out of 100
Rating 5.0 (1)
Status Verified Official
Updated 5 months ago
Version 1.0.0
Models
claudeuniversal
Add to Favorites

Why it matters

Access and discover a vast catalog of AI assets from the Entire VC Spark platform. Seamlessly integrate with your AI tools to find, retrieve, and utilize a wide range of prompts, skills, and models.

Outcomes

What it gets done

01

Search the Spark catalog by query, type, or domain.

02

Retrieve detailed information about specific AI assets using their slugs.

03

Access the raw content of prompts and skills.

04

Discover top-performing assets based on download counts.

Install

Add it to your toolbox

Run in your project directory:

curl -fsSL https://spark.entire.vc/get/spark-mcp-server | bash

Reviews

What people say after installing

5.0
1 ratings

Capabilities

Tools your agent gets

search_assets

Search the catalog by query, type, domain

get_asset

Get full asset details by slug

get_asset_content

Get raw prompt/skill content

list_popular

Top assets by downloads

list_categories

Available domains and LLM tags

Overview

Spark MCP Server

Spark MCP Server exposes search, asset-detail, and raw-content tools over the Entire VC Spark AI-asset catalog, available as a hosted remote endpoint or a local stdio server. Use it when an AI assistant needs to search, browse, or pull content from the Spark catalog directly inside a conversation.

What it does

Spark MCP Server is a Model Context Protocol server for discovering AI assets from the Entire VC Spark catalog (spark.entire.vc). It exposes tools to search the catalog by query, type, or domain, retrieve full details for a specific asset by slug, pull an asset's raw prompt or skill content, list the most popular assets by downloads, and list the available domains and LLM tags. Two resource URIs expose asset content and per-type catalog listings directly as markdown.

When to use - and when NOT to

Use this when an AI assistant needs to search, browse, or pull the actual content of an asset in the Spark catalog - finding a relevant prompt, skill, or agent by keyword or category, or retrieving its raw content to use directly - without leaving the conversation to browse the Spark website. It is scoped to the Spark catalog specifically; paid content requires a Bearer token, so without one, tools that touch gated assets will not return their full content.

Inputs and outputs

Tools: search_assets (query, type, domain filters), get_asset (full details by slug), get_asset_content (raw prompt/skill content), list_popular (top assets by downloads), and list_categories (available domains and LLM tags). Resources: spark://assets/{slug} (asset content as markdown) and spark://catalog/{type} (asset list by type).

Integrations

Two connection modes: a remote hosted server requiring no installation, configured in Claude Desktop or Claude Code by pointing at https://spark.entire.vc/mcp/ (optionally with an Authorization: Bearer <jwt> header for paid-content access), or a local stdio server built from the mcp-server directory with npm install and npm run build, then run via a node command pointing at the built dist/index.js. Local mode is configured through SPARK_API_URL (default https://spark.entire.vc/api/v1) and SPARK_SITE_URL (default https://spark.entire.vc) environment variables.

Who it's for

AI assistant users and developers who want to search, browse, and pull content directly from the Spark AI-asset catalog inside their MCP client instead of visiting the Spark website.

{
  "mcpServers": {
    "spark": {
      "url": "https://spark.entire.vc/mcp/"
    }
  }
}
Source README

Spark MCP Server

MCP (Model Context Protocol) server for discovering AI assets from the Entire VC Spark .

Tools

Tool Description
search_assets Search the catalog by query, type, domain
get_asset Get full asset details by slug
get_asset_content Get raw prompt/skill content
list_popular Top assets by downloads
list_categories Available domains and LLM tags

Resources

URI Description
spark://assets/{slug} Asset content as markdown
spark://catalog/{type} Asset list by type

Connection Options

Option 1: Remote (recommended)

No installation required. Connect directly to the hosted MCP server:

Claude Desktop (~/Library/Application Support/Claude/claude_desktop_config.json):

{
  "mcpServers": {
    "spark": {
      "url": "https://spark.entire.vc/mcp/"
    }
  }
}

Claude Code (.claude/settings.json):

{
  "mcpServers": {
    "spark": {
      "url": "https://spark.entire.vc/mcp/"
    }
  }
}

To enable paid content access, add your Bearer token:

{
  "mcpServers": {
    "spark": {
      "url": "https://spark.entire.vc/mcp/",
      "headers": {
        "Authorization": "Bearer <your-jwt-token>"
      }
    }
  }
}

Option 2: Local (stdio)

Install and run locally via Node.js:

cd mcp-server
npm install
npm run build

Claude Desktop:

{
  "mcpServers": {
    "spark": {
      "command": "node",
      "args": ["/absolute/path/to/evc-spark/mcp-server/dist/index.js"]
    }
  }
}

Claude Code:

{
  "mcpServers": {
    "spark": {
      "command": "node",
      "args": ["/absolute/path/to/evc-spark/mcp-server/dist/index.js"]
    }
  }
}

Environment Variables (local mode only)

Variable Default Description
SPARK_API_URL https://spark.entire.vc/api/v1 Spark API base URL
SPARK_SITE_URL https://spark.entire.vc Spark site URL (for links)

FAQ

Common questions

Trust

How it checks out

Verified 1 rating
Official By maintainer
Downloads 90

Discussion

Questions & comments · 0

Sign In Sign in to leave a comment.