Discover Awesome MCP Servers

JSer.info MCP Server

A Model Context Protocol server that provides search and retrieval capabilities for JSer.info's JavaScript resource database, enabling access to items, posts, product information, and timeline data through various specialized tools.

Plane MCP Server

A Model Context Protocol (MCP) server that enables LLMs to interact with Plane.so, allowing them to manage projects and issues through Plane's API. Using this server, LLMs like Claude can directly interact with your project management workflows while maintaining user control and security.

MCP

Okay, I understand. You want to set up your MCP (presumably a custom application or system) to display company information and stock prices, leveraging the Claude model (likely for natural language processing or data retrieval). Here's a breakdown of the steps and considerations, along with a conceptual outline. Keep in mind that the specific implementation will depend heavily on the architecture of your MCP system and the available APIs for Claude and financial data. **Conceptual Outline** 1. **Data Sources:** * **Company Information:** Identify a reliable source of company information. This could be a database you maintain, a commercial API (e.g., Crunchbase, Clearbit), or a web scraping solution (less reliable but sometimes necessary). * **Stock Prices:** Choose a stock market data provider. Popular options include: * **IEX Cloud:** Relatively affordable and easy to use. * **Alpha Vantage:** Offers a free tier with limitations. * **Financial Modeling Prep:** Another option with various data points. * **Refinitiv (formerly Thomson Reuters):** More expensive but provides comprehensive data. * **Bloomberg:** High-end, professional-grade data. * **Claude API:** You'll need access to the Claude API. This usually involves signing up for an account and obtaining an API key. 2. **MCP System Integration:** * **API Endpoints:** Design API endpoints within your MCP system to handle requests for company information and stock prices. For example: * `/company/{company_name}`: Returns company details. * `/stock/{ticker_symbol}`: Returns stock price information. * **Data Storage (Optional):** Consider caching frequently accessed data in a database within your MCP system to reduce API calls to external services and improve response times. 3. **Claude Integration:** * **Natural Language Understanding (NLU):** Use Claude to understand user queries. For example, if a user types "Tell me about Apple's stock," Claude should identify the intent ("get stock information") and the entity ("Apple"). * **Data Retrieval:** Based on the NLU results, construct the appropriate API calls to your data sources (company information and stock price APIs). * **Response Formatting:** Use Claude to format the retrieved data into a human-readable response. This is where Claude's natural language generation (NLG) capabilities shine. 4. **Error Handling:** * Implement robust error handling to gracefully handle API failures, invalid user input, and other potential issues. **Detailed Steps and Considerations** 1. **Choose Your Data Sources:** * **Company Information:** * **API:** If you choose an API, review its documentation to understand how to query for company information (e.g., by name, ticker symbol, or domain). Consider the API's rate limits and pricing. * **Database:** If you have a database, ensure it's well-structured and indexed for efficient querying. * **Web Scraping:** Use a library like Beautiful Soup (Python) or Cheerio (Node.js) to scrape data from websites. Be aware of the legal and ethical implications of web scraping. Websites can change their structure, breaking your scraper. * **Stock Prices:** * **API:** Select a stock market data API based on your budget and data requirements. Pay attention to: * **Real-time vs. Delayed Data:** Real-time data is more expensive. Delayed data (e.g., 15-minute delay) may be sufficient for some use cases. * **Historical Data:** If you need historical stock prices, ensure the API provides it. * **API Limits:** Understand the API's rate limits and usage restrictions. 2. **Set Up Your MCP API Endpoints:** * Use a framework like Flask (Python), Express.js (Node.js), or Spring Boot (Java) to create your API endpoints. * Implement authentication and authorization if necessary to protect your API. * Use a consistent API design (e.g., RESTful principles). 3. **Integrate with Claude:** * **Install the Claude API client:** Follow the instructions in the Claude API documentation to install the appropriate client library for your programming language. * **Authentication:** Use your Claude API key to authenticate your requests. * **NLU (Natural Language Understanding):** * Send the user's query to the Claude API. * Use Claude's NLU capabilities to extract the intent and entities. You might need to fine-tune Claude for your specific use case. Consider using prompt engineering to guide Claude's understanding. For example: ```python import anthropic client = anthropic.Anthropic(api_key="YOUR_ANTHROPIC_API_KEY") def get_intent_and_entities(query): prompt = f"""You are a helpful assistant that analyzes user queries to determine their intent and extract relevant entities. User Query: {query} Intent: (The user's goal, e.g., "get stock information", "get company details") Entities: (A list of relevant entities, e.g., "Apple", "AAPL") """ response = client.completions.create( model="claude-v1.3", # Or the latest Claude model prompt=prompt, max_tokens_to_sample=200, ) # Parse the response to extract the intent and entities. This will require some string manipulation. # You might want to use regular expressions to make this easier. # Example: # intent = response.completion.split("Intent: ")[1].split("\n")[0].strip() # entities = response.completion.split("Entities: ")[1].split("\n")[0].strip().split(", ") # Return the intent and entities return intent, entities ``` * **Data Retrieval:** * Based on the extracted intent and entities, construct the appropriate API calls to your data sources. * Handle potential errors from the data source APIs. * **Response Formatting (NLG - Natural Language Generation):** * Use Claude to format the retrieved data into a human-readable response. Again, prompt engineering is crucial here. For example: ```python def format_response(company_name, stock_price, company_description): prompt = f"""You are a helpful assistant that formats company information and stock prices into a human-readable response. Company Name: {company_name} Stock Price: {stock_price} Company Description: {company_description} Response: """ response = client.completions.create( model="claude-v1.3", # Or the latest Claude model prompt=prompt, max_tokens_to_sample=300, ) return response.completion ``` 4. **Error Handling:** * Implement `try...except` blocks (Python) or similar error handling mechanisms in your code. * Log errors to a file or monitoring system for debugging. * Return informative error messages to the user. **Example Code Snippet (Conceptual - Python with Flask)** ```python from flask import Flask, request, jsonify import anthropic import requests # For making API calls to data sources app = Flask(__name__) # Replace with your actual API keys and data source URLs CLAUDE_API_KEY = "YOUR_CLAUDE_API_KEY" STOCK_API_URL = "https://api.example.com/stock/{ticker}" # Example COMPANY_API_URL = "https://api.example.com/company/{name}" # Example client = anthropic.Anthropic(api_key=CLAUDE_API_KEY) def get_intent_and_entities(query): # (Implementation as described above) # This is a placeholder. You'll need to parse the Claude response. if "stock" in query.lower(): intent = "get stock information" entity = query.split("stock of ")[-1].strip() # Very basic example elif "company" in query.lower(): intent = "get company information" entity = query.split("about ")[-1].strip() # Very basic example else: intent = "unknown" entity = None return intent, entity def get_stock_price(ticker): try: response = requests.get(STOCK_API_URL.format(ticker=ticker)) response.raise_for_status() # Raise HTTPError for bad responses (4xx or 5xx) data = response.json() return data["price"] except requests.exceptions.RequestException as e: print(f"Error fetching stock price: {e}") return None def get_company_description(company_name): try: response = requests.get(COMPANY_API_URL.format(name=company_name)) response.raise_for_status() data = response.json() return data["description"] except requests.exceptions.RequestException as e: print(f"Error fetching company description: {e}") return None def format_response(company_name, stock_price, company_description): # (Implementation as described above) prompt = f"""You are a helpful assistant that formats company information and stock prices into a human-readable response. Company Name: {company_name} Stock Price: {stock_price} Company Description: {company_description} Response: """ response = client.completions.create( model="claude-v1.3", # Or the latest Claude model prompt=prompt, max_tokens_to_sample=300, ) return response.completion @app.route("/query", methods=["POST"]) def handle_query(): query = request.json.get("query") if not query: return jsonify({"error": "Missing query parameter"}), 400 intent, entity = get_intent_and_entities(query) if intent == "get stock information": stock_price = get_stock_price(entity) if stock_price is None: return jsonify({"error": "Could not retrieve stock price"}), 500 company_description = get_company_description(entity) # Try to get company description too if company_description is None: company_description = "No description available." formatted_response = format_response(entity, stock_price, company_description) return jsonify({"response": formatted_response}) elif intent == "get company information": company_description = get_company_description(entity) if company_description is None: return jsonify({"error": "Could not retrieve company information"}), 500 formatted_response = format_response(entity, "N/A", company_description) # Stock price not applicable return jsonify({"response": formatted_response}) else: return jsonify({"error": "Unknown intent"}), 400 if __name__ == "__main__": app.run(debug=True) ``` **Key Improvements and Considerations:** * **Prompt Engineering:** The quality of Claude's responses depends heavily on the prompts you provide. Experiment with different prompts to optimize the results. Consider using few-shot learning (providing examples in the prompt) to guide Claude's behavior. * **Error Handling:** The example code includes basic error handling, but you should implement more robust error handling in a production environment. * **API Rate Limits:** Be mindful of the API rate limits of your data sources and the Claude API. Implement caching and rate limiting in your MCP system to avoid exceeding these limits. * **Security:** Protect your API keys and other sensitive information. Use environment variables to store API keys and avoid hardcoding them in your code. * **Scalability:** Consider the scalability of your MCP system. If you expect a large number of requests, you may need to use a load balancer and multiple servers. * **Testing:** Thoroughly test your MCP system to ensure it is working correctly. Write unit tests and integration tests to verify the functionality of your code. * **Monitoring:** Monitor your MCP system to identify and resolve issues quickly. Use a monitoring tool to track API response times, error rates, and other metrics. * **Data Validation:** Validate the data returned from the APIs to ensure it is accurate and consistent. * **Asynchronous Operations:** For long-running tasks (like calling external APIs), consider using asynchronous operations (e.g., with `asyncio` in Python) to improve the responsiveness of your MCP system. * **User Interface:** Consider how users will interact with your MCP system. You may want to create a web interface or a command-line interface. **Summary** This is a complex project that requires careful planning and execution. Start with a simple prototype and gradually add more features. Focus on the core functionality first (data retrieval and response formatting) and then add more advanced features like error handling, caching, and rate limiting. Remember to consult the documentation for the Claude API and your chosen data source APIs. Good luck!

MCP + DeepSeek AI Integration Server

A Node.js-based FastMCP protocol server that integrates DeepSeek AI capabilities for intelligent conversations and code analysis, providing tool invocation abilities through the MCP protocol.

FastAPI MCP Server

Wraps FastAPI REST endpoints as MCP tools, enabling natural language interaction with user management, task management, and mathematical calculations through Gemini CLI.

Weather MCP Server

Enables AI assistants to access real-time US weather forecasts and alerts through the National Weather Service API.

Karakeep MCP server

Search and add bookmarks

Atlassian Bitbucket MCP

Enables interaction with Bitbucket through the Model Context Protocol, allowing users to manage pull requests, add comments, review code, create tasks, and perform other repository operations using natural language.

SocialGuessSkills

A multi-agent framework that models complex social systems through a coordinated workflow of seven specialized agents covering areas like economics, governance, and risk. It enables users to generate structured, verifiable social system models from basic assumptions using the Model Context Protocol.

CC Fig MCP

Enables bidirectional control between Claude Code and Figma with real-time sync, allowing creation and manipulation of design elements, component management, and document inspection through natural language commands.

MCP Store Greeting Server

Provides personalized store greetings combined with real-time weather information based on store location. Includes an admin web interface for managing stores with address autocomplete and map-based location selection.

AGR MCP Server

Enables querying genomics data from the Alliance of Genome Resources across model organisms including human, mouse, rat, zebrafish, fly, worm, yeast, and xenopus. Supports gene searches, disease associations, expression data, orthologs, phenotypes, and molecular interactions through natural language.

amazon-ads-mcp-server

amazon-ads-mcp-server

Personal Knowledge MCP Server

Indexes local and enterprise documents to provide a unified personal knowledge base for AI clients via the Model Context Protocol. It supports full-text search across various file formats and integrates with platforms like Feishu and WeChat Work.

Mcp Pallete

シンプルな、おもちゃのような Minecraft サーバーで、画像の色を取得して、それをもとに PNG パレットを作成できます。

Jotai MCP Server

Jotai MCPサーバー (Jotai MCP Sābā)

Dexscreener MCP server

Dexscreener API の MCP サーバー - あなたの AI エージェントが Dexscreener の無料でオープンな API を使用して、オンチェーンの価格をチェックできます。

Pixabay MCP Server

Enables searching and retrieving royalty-free images and videos from Pixabay's library with advanced filtering options including category, orientation, color, size, and quality preferences.

Agent Care

EMR (Cerner や Epic など) 上の FHIR データや医療リソースとやり取りするためのヘルスケアツールを提供する MCP サーバー

Scripting Docs MCP

Provides semantic search over ScriptingApp documentation by converting Markdown/MDX files into a LlamaIndex vector store. Supports multi-language indexing and enables natural language queries against technical documentation through MCP tools.

Solana Agent Kit MCP Server

Claude AIにオンチェーンツールを提供するためのモデルコンテキストプロトコルサーバー。資産管理、トークン操作の実行、ネットワーク情報の取得など、Solanaブロックチェーンとのインタラクションを標準化されたインターフェースを通じて可能にする。

MCP Code Checker

MCPサーバーは、コード品質チェック（pylintとpytest）を提供し、分析と修正のためにLLMフレンドリーなスマートなプロンプトを使用します。Claudeや他のAIアシスタントがあなたのコードを分析し、改善点を提案することを可能にします。

Plaid Transactions MCP Server

Enables secure interaction with Plaid's Transactions API to sync, search, and analyze financial transactions from linked bank accounts using natural language queries.

ASCII Banner MCP Server

An MCP server for getting ASCII banners

MaaMCP

Enables AI assistants to automate Android devices and Windows desktop applications through MaaFramework, supporting multi-device control, OCR text recognition, screen capture, clicking, swiping, text input, and Pipeline generation for reusable automation workflows.

Stock Data MCP Server

Provides comprehensive stock market data across US, Hong Kong, and Chinese markets, combining real-time quotes, historical data, fundamentals, and financial statements from multiple sources including Yahoo Finance, Finnhub, Tushare, and Futu OpenAPI.

MCP Sui Tools

Suiブロックチェーンと統合されたツールキットで、ユーザーがウォレットアドレスを提供すると、Claudeがテストネットフォーセットツールを通じてテストトークンをリクエストできるようになります。

SSH Remote MCP Server

Enables SSH remote access to servers through Claude, allowing users to execute commands, transfer files via SFTP, and manage multiple remote connections using natural language.

ynab-mcpb

MCP server for YNAB. Reconcile bank statements, itemize receipts, manage transactions — all through natural language.