MCP Bluesky Stock Sentiment
Analyzes recent Bluesky posts about stocks and provides sentiment analysis using Cloudflare Workers AI. Searches for stock-related content on Bluesky and returns sentiment summaries with confidence scores and post details.
README
MCP Bluesky Stock Sentiment
Analyze recent Bluesky posts about stocks and return a concise sentiment summary using Cloudflare Workers AI — exposed as an MCP server you can connect to from any MCP‑compatible client.
What you get
- Cloudflare Worker with Hono, serving an MCP endpoint at
/mcp. - Tool:
analyze_stock_posts— searches recent Bluesky posts, filters for stock‑related content, and runs sentiment via@cf/huggingface/distilbert-sst-2-int8. - Simple JSON metadata on
/for quick health checks.
Quickstart
- Prereqs: Node 18+, Cloudflare account, Wrangler auth (
npx wrangler login). - Install:
npm install - Dev:
npm run dev(serves onhttp://localhost:8787) - Deploy:
npm run deploy - Typegen:
npm run cf-typegen(optional; syncs Worker bindings types)
MCP Endpoint
- Local:
http://localhost:8787/mcp - Prod:
<your-workers-subdomain>.workers.dev/mcp - Server name:
bluesky-stock-sentiment(seesrc/index.ts).
Available Tool
analyze_stock_posts- Input:
{ count?: number }— number of posts to analyze (1–10, default 2) - Behavior: Fetches recent posts from Bluesky’s public search API, filters to stock‑related text, analyzes each with Workers AI, and returns a textual summary (totals, percentages, average confidence, and per‑post details with author, likes, reposts, timestamp, and URI).
- Input:
Config and Extensibility
- Sentiment model:
src/lib/sentiment.ts- Default:
@cf/huggingface/distilbert-sst-2-int8 - Swap to any Workers AI text‑classification model if you prefer.
- Default:
- Stock filtering:
src/lib/stock-filter.ts- Update the keyword list to fine‑tune what counts as a “stock post”.
- Bluesky search:
src/lib/bluesky-api.ts- Uses the public unauthenticated search API and basic heuristics to avoid duplicates.
- Firehose (optional):
src/lib/firehose.ts- Example collector that connects to Bluesky’s firehose via WebSocket and filters in real time. This is not wired into the Worker route; use it as a reference for server/runtime environments that support outbound WebSocket connections.
Endpoints
GET /— returns service metadata (name, version, description, available tools).ALL /mcp— MCP over HTTP transport (viamcp-lite). Point your MCP client here.
Notes
- You must bind Workers AI in Wrangler (already configured in
wrangler.jsoncunderai.binding = "AI"). - Bluesky’s public search API may rate‑limit or change behavior; the tool is resilient but results vary over time.
Project Layout
src/index.ts— Hono app, MCP server wiring, routes.src/mcp/tools.ts—analyze_stock_poststool registration and handler.src/lib/*— sentiment analysis, filtering, Bluesky integrations.
License
- Provided as‑is; add a license if you plan to distribute.
推荐服务器
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。
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。
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。
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。
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。
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。
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。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
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。