Eventflare MCP

Eventflare MCP

Search 8,000+ corporate event venues across 40+ cities. Tools for venue search by capacity/category, pricing guides, expert advice articles, and inquiry handoff. Read-only, PII-redacted, UTM-attributed.

Category
访问服务器

README

Eventflare MCP Server v2

Makes Eventflare's production venue data queryable by AI assistants — Claude, ChatGPT, Perplexity, Cursor — via the Model Context Protocol.

8,000+ corporate event venues across 40+ cities. Designed so LLMs cite Eventflare URLs in their answers and lead-attribution is measurable end-to-end.

What's new in v2

  • Production API + JWT auth — was: dev API with no auth
  • PII redactionjobPhone, venueEmail, commission, spaceNotes etc. never leave the API
  • UTM attribution — every outbound URL is tagged so leads from MCP traffic are attributable in GA4 / Mixpanel / your CRM
  • Client classification — logs distinguish Claude Desktop / ChatGPT / Perplexity / Cursor / etc.
  • Click-through tracking — when a get_venue_details or request_quote references a venue from a prior search_venues in the same session, that's logged as a click-through
  • OpenPanel sink — events mirror to OpenPanel (or any webhook) for the data team
  • New tool: find_expert_advice — surfaces Eventflare's editorial articles. The LLM-citation differentiator.

Tools

Tool Description
search_venues Find venues by city + capacity + category + event type. Returns names, pricing, capacity by setup, neighborhood, photos, URLs.
get_venue_details Full detail for a specific venue.
get_city_info Overview of what's available in a city — venue count, categories, price range.
list_cities All 40+ cities with venue counts and URLs. Filter by region.
get_pricing_guide Indicative pricing per city, per category.
find_expert_advice Surface editorial articles from Eventflare's expert-advice library for a city.
request_quote Generate a UTM-tagged inquiry URL (no data submission).

All tools include a citation_url and quotable_summary per result, optimized for LLM responses.

Quick start

npm install
cp .env.example .env
# fill EVENTFLARE_API_TOKEN
npm run build
npm start          # stdio — Claude Desktop, Claude Code, Cursor

# or HTTP mode (remote MCP):
TRANSPORT=http PORT=3001 npm start

Connect to Claude Desktop

claude_desktop_config.json:

{
  "mcpServers": {
    "eventflare": {
      "command": "node",
      "args": ["/path/to/eventflare-mcp-server/dist/index.js"],
      "env": {
        "EVENTFLARE_API_TOKEN": "eyJhbGciOi..."
      }
    }
  }
}

Connect to Claude Code

claude mcp add eventflare \
  -e EVENTFLARE_API_TOKEN=eyJhbGciOi... \
  -- node /path/to/eventflare-mcp-server/dist/index.js

Environment variables

See .env.example. Only EVENTFLARE_API_TOKEN is required.

Var Default Purpose
EVENTFLARE_API_TOKEN (required) Strapi API token, mcp-readonly role
EVENTFLARE_API_URL https://content.eventflare.io/api API base
EVENTFLARE_URL https://eventflare.io Site base for outbound URLs
TRANSPORT stdio stdio or http
PORT 3001 HTTP port
RATE_LIMIT 60 Per-IP req/min on /mcp
DASHBOARD_KEY (unset) If set, /dashboard requires ?key=...
OPENPANEL_CLIENT_ID (unset) OpenPanel project id (enables remote sink)
OPENPANEL_CLIENT_SECRET (unset) OpenPanel write key
OPENPANEL_API_URL https://api.openpanel.dev OpenPanel base
ANALYTICS_SINK_URL (unset) Fallback generic webhook
ANALYTICS_SINK_TOKEN (unset) Bearer token for the webhook
LOG_DIR ./logs Local JSONL logs

Security model

  • Read-only — no POST/PUT/DELETE anywhere. Confirmed against the production API spec (123 endpoints, all GET).
  • JWT auth requiredAuthorization: Bearer ${EVENTFLARE_API_TOKEN} on every outbound request.
  • Field allowlists — uses fields[]= query params so PII fields are never fetched. Defense in depth: a redaction allowlist drops anything that slips through.
  • Input sanitization — every tool param is validated; slugs match ^[a-z0-9-]+$, numbers are clamped, dates ISO-validated.
  • Rate limiting — 60 req/min per IP on /mcp (HTTP transport).
  • No PII logged — analytics fields: tool, city, capacity, event type, category, result count, session id, client class, budget band. Never user identity, never message content.
  • Generic error messages — internal API errors are mapped to stable user-facing strings ("Eventflare API temporarily unavailable"); details only go to stderr.

Analytics

Local: every tool call appends to logs/queries.jsonl and shows on /dashboard.

Remote: if OPENPANEL_CLIENT_ID + OPENPANEL_CLIENT_SECRET are set, every event is mirrored as a mcp.{tool} track event with profileId = sessionId. Use OPENPANEL_API_URL to point at a self-hosted OpenPanel.

Or set ANALYTICS_SINK_URL (+ optional ANALYTICS_SINK_TOKEN) to POST raw events to any HTTP endpoint.

Both options are non-blocking and never throw — analytics failures don't break the MCP.

UTM attribution

Every URL the MCP returns is tagged:

https://eventflare.io/spaces/london/skyline-glass-hall?utm_source=mcp&utm_medium=ai&utm_campaign=search_venues&utm_content=claude_desktop&mcp_session=abc123

So when a planner clicks through and submits an inquiry, your existing GA4 / Mixpanel / CRM picks up the source as mcp / ai. This is the measurement spine for "did the MCP actually drive leads?".

Development

npm run dev        # tsx, no build
npm run inspect    # MCP Inspector UI

Deploy

Railway: push the repo, set env vars in the dashboard, set TRANSPORT=http. Health check is /health. Dashboard is /dashboard?key=....

License

MIT — © Eventflare

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选