mcp-use 0.1.20 → 0.2.0

package/dist/examples/ai_sdk_example.d.ts

@@ -1,23 +0,0 @@
-/**
- * AI SDK Integration Example
- *
- * This example demonstrates how to use MCPAgent's streamEvents() method
- * with Vercel AI SDK's LangChainAdapter for building streaming UIs.
- *
- * This pattern is useful for:
- * - Next.js API routes with useCompletion/useChat hooks
- * - Real-time streaming applications
- * - Building chat interfaces with token-by-token updates
- */
-import type { StreamEvent } from '../index.js';
-declare function streamEventsToAISDK(streamEvents: AsyncGenerator<StreamEvent, void, void>): AsyncGenerator<string, void, void>;
-declare function createReadableStreamFromGenerator(generator: AsyncGenerator<string, void, void>): ReadableStream<string>;
-declare function streamEventsToAISDKWithTools(streamEvents: AsyncGenerator<StreamEvent, void, void>): AsyncGenerator<string, void, void>;
-declare function createApiHandler(): Promise<(request: {
-    prompt: string;
-}) => Promise<Response>>;
-declare function createEnhancedApiHandler(): Promise<(request: {
-    prompt: string;
-}) => Promise<Response>>;
-export { createApiHandler, createEnhancedApiHandler, createReadableStreamFromGenerator, streamEventsToAISDK, streamEventsToAISDKWithTools, };
-//# sourceMappingURL=ai_sdk_example.d.ts.map

package/dist/examples/ai_sdk_example.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"ai_sdk_example.d.ts","sourceRoot":"","sources":["../../examples/ai_sdk_example.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AAEH,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,aAAa,CAAA;AAU9C,iBAAgB,mBAAmB,CACjC,YAAY,EAAE,cAAc,CAAC,WAAW,EAAE,IAAI,EAAE,IAAI,CAAC,GACpD,cAAc,CAAC,MAAM,EAAE,IAAI,EAAE,IAAI,CAAC,CAUpC;AAGD,iBAAS,iCAAiC,CACxC,SAAS,EAAE,cAAc,CAAC,MAAM,EAAE,IAAI,EAAE,IAAI,CAAC,GAC5C,cAAc,CAAC,MAAM,CAAC,CAcxB;AAGD,iBAAgB,4BAA4B,CAC1C,YAAY,EAAE,cAAc,CAAC,WAAW,EAAE,IAAI,EAAE,IAAI,CAAC,GACpD,cAAc,CAAC,MAAM,EAAE,IAAI,EAAE,IAAI,CAAC,CAqBpC;AAGD,iBAAe,gBAAgB,sBAwBM;IAAE,MAAM,EAAE,MAAM,CAAA;CAAE,wBAsBtD;AAGD,iBAAe,wBAAwB,sBAuBM;IAAE,MAAM,EAAE,MAAM,CAAA;CAAE,wBAkB9D;AAoFD,OAAO,EACL,gBAAgB,EAChB,wBAAwB,EACxB,iCAAiC,EACjC,mBAAmB,EACnB,4BAA4B,GAC7B,CAAA"}

package/dist/examples/ai_sdk_example.js

@@ -1,213 +0,0 @@
-/**
- * AI SDK Integration Example
- *
- * This example demonstrates how to use MCPAgent's streamEvents() method
- * with Vercel AI SDK's LangChainAdapter for building streaming UIs.
- *
- * This pattern is useful for:
- * - Next.js API routes with useCompletion/useChat hooks
- * - Real-time streaming applications
- * - Building chat interfaces with token-by-token updates
- */
-import { ChatAnthropic } from '@langchain/anthropic';
-import { LangChainAdapter } from 'ai';
-import { config } from 'dotenv';
-import { MCPAgent, MCPClient } from '../index.js';
-// Load environment variables
-config();
-// Utility function to convert streamEvents to AI SDK compatible stream
-async function* streamEventsToAISDK(streamEvents) {
-    for await (const event of streamEvents) {
-        // Only yield the actual content tokens from chat model streams
-        if (event.event === 'on_chat_model_stream' && event.data?.chunk?.text) {
-            const textContent = event.data.chunk.text;
-            if (typeof textContent === 'string' && textContent.length > 0) {
-                yield textContent;
-            }
-        }
-    }
-}
-// Convert async generator to ReadableStream for AI SDK compatibility
-function createReadableStreamFromGenerator(generator) {
-    return new ReadableStream({
-        async start(controller) {
-            try {
-                for await (const chunk of generator) {
-                    controller.enqueue(chunk);
-                }
-                controller.close();
-            }
-            catch (error) {
-                controller.error(error);
-            }
-        },
-    });
-}
-// Enhanced adapter that includes tool information
-async function* streamEventsToAISDKWithTools(streamEvents) {
-    for await (const event of streamEvents) {
-        switch (event.event) {
-            case 'on_chat_model_stream':
-                if (event.data?.chunk?.text) {
-                    const textContent = event.data.chunk.text;
-                    if (typeof textContent === 'string' && textContent.length > 0) {
-                        yield textContent;
-                    }
-                }
-                break;
-            case 'on_tool_start':
-                yield `\n🔧 Using tool: ${event.name}\n`;
-                break;
-            case 'on_tool_end':
-                yield `\n✅ Tool completed: ${event.name}\n`;
-                break;
-        }
-    }
-}
-// Example: Basic AI SDK API route handler
-async function createApiHandler() {
-    const everythingServer = {
-        mcpServers: {
-            everything: {
-                command: 'npx',
-                args: ['-y', '@modelcontextprotocol/server-everything'],
-            },
-        },
-    };
-    const client = new MCPClient(everythingServer);
-    const llm = new ChatAnthropic({
-        model: 'claude-sonnet-4-20250514',
-        temperature: 0.1,
-    });
-    const agent = new MCPAgent({
-        llm,
-        client,
-        maxSteps: 5,
-        verbose: false,
-    });
-    // Simulate an API route handler
-    const apiHandler = async (request) => {
-        try {
-            // Get streamEvents from MCPAgent
-            const streamEvents = agent.streamEvents(request.prompt);
-            // Convert to AI SDK compatible format
-            const aiSDKStream = streamEventsToAISDK(streamEvents);
-            const readableStream = createReadableStreamFromGenerator(aiSDKStream);
-            // Use LangChainAdapter to create a Response compatible with AI SDK
-            return LangChainAdapter.toDataStreamResponse(readableStream);
-        }
-        catch (error) {
-            console.error('Error in API handler:', error);
-            throw error;
-        }
-        finally {
-            await client.closeAllSessions();
-        }
-    };
-    return apiHandler;
-}
-// Example: Enhanced API handler with tool visibility
-async function createEnhancedApiHandler() {
-    const everythingServer = {
-        mcpServers: {
-            everything: {
-                command: 'npx',
-                args: ['-y', '@modelcontextprotocol/server-everything'],
-            },
-        },
-    };
-    const client = new MCPClient(everythingServer);
-    const llm = new ChatAnthropic({
-        model: 'claude-sonnet-4-20250514',
-        temperature: 0.1,
-    });
-    const agent = new MCPAgent({
-        llm,
-        client,
-        maxSteps: 8,
-        verbose: false,
-    });
-    const enhancedApiHandler = async (request) => {
-        try {
-            const streamEvents = agent.streamEvents(request.prompt);
-            const enhancedStream = streamEventsToAISDKWithTools(streamEvents);
-            const readableStream = createReadableStreamFromGenerator(enhancedStream);
-            return LangChainAdapter.toDataStreamResponse(readableStream);
-        }
-        catch (error) {
-            console.error('Error in enhanced API handler:', error);
-            throw error;
-        }
-        finally {
-            await client.closeAllSessions();
-        }
-    };
-    return enhancedApiHandler;
-}
-// Example: Simulated Next.js API route
-async function simulateNextJSApiRoute() {
-    console.log('🚀 Simulating Next.js API Route with AI SDK Integration\n');
-    const apiHandler = await createApiHandler();
-    // Simulate a request
-    const request = {
-        prompt: 'What\'s the current time? Also, list the files in the current directory.',
-    };
-    console.log(`📝 Request: ${request.prompt}\n`);
-    console.log('📡 Streaming response:\n');
-    try {
-        const response = await apiHandler(request);
-        if (response.body) {
-            const reader = response.body.getReader();
-            const decoder = new TextDecoder();
-            while (true) {
-                const { done, value } = await reader.read();
-                if (done)
-                    break;
-                const chunk = decoder.decode(value);
-                process.stdout.write(chunk);
-            }
-        }
-    }
-    catch (error) {
-        console.error('❌ Error:', error);
-    }
-    console.log('\n\n✅ API Route simulation complete');
-}
-// Example: Enhanced streaming with tool visibility
-async function simulateEnhancedStreaming() {
-    console.log('\n\n🚀 Enhanced Streaming with Tool Visibility\n');
-    const enhancedHandler = await createEnhancedApiHandler();
-    const request = {
-        prompt: 'Check the current time and create a file with a timestamp. Then tell me what tools you used.',
-    };
-    console.log(`📝 Request: ${request.prompt}\n`);
-    console.log('📡 Enhanced streaming response:\n');
-    try {
-        const response = await enhancedHandler(request);
-        if (response.body) {
-            const reader = response.body.getReader();
-            const decoder = new TextDecoder();
-            while (true) {
-                const { done, value } = await reader.read();
-                if (done)
-                    break;
-                const chunk = decoder.decode(value);
-                process.stdout.write(chunk);
-            }
-        }
-    }
-    catch (error) {
-        console.error('❌ Error:', error);
-    }
-    console.log('\n\n✅ Enhanced streaming complete');
-}
-// Run all examples
-async function runAllExamples() {
-    await simulateNextJSApiRoute();
-    await simulateEnhancedStreaming();
-}
-// Export utilities for reuse
-export { createApiHandler, createEnhancedApiHandler, createReadableStreamFromGenerator, streamEventsToAISDK, streamEventsToAISDKWithTools, };
-if (import.meta.url === `file://${process.argv[1]}`) {
-    runAllExamples().catch(console.error);
-}

package/dist/examples/airbnb_use.d.ts

@@ -1,10 +0,0 @@
-/**
- * Example demonstrating how to use mcp-use with Airbnb.
- *
- * This example shows how to connect an LLM to Airbnb through MCP tools
- * to perform tasks like searching for accommodations.
- *
- * Special Thanks to https://github.com/openbnb-org/mcp-server-airbnb for the server.
- */
-export {};
-//# sourceMappingURL=airbnb_use.d.ts.map

package/dist/examples/airbnb_use.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"airbnb_use.d.ts","sourceRoot":"","sources":["../../examples/airbnb_use.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG"}

package/dist/examples/airbnb_use.js

@@ -1,43 +0,0 @@
-/**
- * Example demonstrating how to use mcp-use with Airbnb.
- *
- * This example shows how to connect an LLM to Airbnb through MCP tools
- * to perform tasks like searching for accommodations.
- *
- * Special Thanks to https://github.com/openbnb-org/mcp-server-airbnb for the server.
- */
-import { ChatOpenAI } from '@langchain/openai';
-import { config } from 'dotenv';
-import { MCPAgent, MCPClient } from '../index.js';
-// Load environment variables from .env file
-config();
-async function runAirbnbExample() {
-    // Create MCPClient with Airbnb configuration
-    const config = {
-        mcpServers: {
-            airbnb: {
-                command: 'npx',
-                args: ['-y', '@openbnb/mcp-server-airbnb', '--ignore-robots-txt'],
-            },
-        },
-    };
-    const client = new MCPClient(config);
-    // Create LLM - you can choose between different models
-    const llm = new ChatOpenAI({ model: 'gpt-4o' });
-    // Create agent with the client
-    const agent = new MCPAgent({ llm, client, maxSteps: 30 });
-    try {
-        // Run a query to search for accommodations
-        const result = await agent.run('Find me a nice place to stay in Barcelona for 2 adults '
-            + 'for a week in August. I prefer places with a pool and '
-            + 'good reviews. Show me the top 3 options.', 30);
-        console.error(`\nResult: ${result}`);
-    }
-    finally {
-        // Ensure we clean up resources properly
-        await client.closeAllSessions();
-    }
-}
-if (import.meta.url === `file://${process.argv[1]}`) {
-    runAirbnbExample().catch(console.error);
-}

package/dist/examples/blender_use.d.ts

@@ -1,15 +0,0 @@
-/**
- * Blender MCP example for mcp-use.
- *
- * This example demonstrates how to use the mcp-use library with MCPClient
- * to connect an LLM to Blender through MCP tools via WebSocket.
- * The example assumes you have installed the Blender MCP addon from:
- * https://github.com/ahujasid/blender-mcp
- *
- * Make sure the addon is enabled in Blender preferences and the WebSocket
- * server is running before executing this script.
- *
- * Special thanks to https://github.com/ahujasid/blender-mcp for the server.
- */
-export {};
-//# sourceMappingURL=blender_use.d.ts.map

package/dist/examples/blender_use.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"blender_use.d.ts","sourceRoot":"","sources":["../../examples/blender_use.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;GAYG"}

package/dist/examples/blender_use.js

@@ -1,39 +0,0 @@
-/**
- * Blender MCP example for mcp-use.
- *
- * This example demonstrates how to use the mcp-use library with MCPClient
- * to connect an LLM to Blender through MCP tools via WebSocket.
- * The example assumes you have installed the Blender MCP addon from:
- * https://github.com/ahujasid/blender-mcp
- *
- * Make sure the addon is enabled in Blender preferences and the WebSocket
- * server is running before executing this script.
- *
- * Special thanks to https://github.com/ahujasid/blender-mcp for the server.
- */
-import { ChatAnthropic } from '@langchain/anthropic';
-import { config } from 'dotenv';
-import { MCPAgent, MCPClient } from '../index.js';
-// Load environment variables from .env file
-config();
-async function runBlenderExample() {
-    // Create MCPClient with Blender MCP configuration
-    const config = { mcpServers: { blender: { command: 'uvx', args: ['blender-mcp'] } } };
-    const client = MCPClient.fromDict(config);
-    // Create LLM
-    const llm = new ChatAnthropic({ model: 'claude-3-5-sonnet-20240620' });
-    // Create agent with the client
-    const agent = new MCPAgent({ llm, client, maxSteps: 30 });
-    try {
-        // Run the query
-        const result = await agent.run('Create an inflatable cube with soft material and a plane as ground.', 30);
-        console.error(`\nResult: ${result}`);
-    }
-    finally {
-        // Ensure we clean up resources properly
-        await client.closeAllSessions();
-    }
-}
-if (import.meta.url === `file://${process.argv[1]}`) {
-    runBlenderExample().catch(console.error);
-}

package/dist/examples/browser_use.d.ts

@@ -1,10 +0,0 @@
-/**
- * Basic usage example for mcp-use.
- *
- * This example demonstrates how to use the mcp-use library with MCPClient
- * to connect any LLM to MCP tools through a unified interface.
- *
- * Special thanks to https://github.com/microsoft/playwright-mcp for the server.
- */
-export {};
-//# sourceMappingURL=browser_use.d.ts.map

package/dist/examples/browser_use.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"browser_use.d.ts","sourceRoot":"","sources":["../../examples/browser_use.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG"}

package/dist/examples/browser_use.js

@@ -1,46 +0,0 @@
-/**
- * Basic usage example for mcp-use.
- *
- * This example demonstrates how to use the mcp-use library with MCPClient
- * to connect any LLM to MCP tools through a unified interface.
- *
- * Special thanks to https://github.com/microsoft/playwright-mcp for the server.
- */
-import path from 'node:path';
-import { fileURLToPath } from 'node:url';
-import { ChatOpenAI } from '@langchain/openai';
-import { config } from 'dotenv';
-import { MCPAgent, MCPClient } from '../index.js';
-// Load environment variables from .env file
-config();
-const __filename = fileURLToPath(import.meta.url);
-const __dirname = path.dirname(__filename);
-async function main() {
-    const config = {
-        mcpServers: {
-            playwright: {
-                command: 'npx',
-                args: ['@playwright/mcp@latest'],
-                env: {
-                    DISPLAY: ':1',
-                },
-            },
-        },
-    };
-    // Create MCPClient from config file
-    const client = new MCPClient(config);
-    // Create LLM
-    const llm = new ChatOpenAI({ model: 'gpt-4o' });
-    // const llm = init_chat_model({ model: "llama-3.1-8b-instant", model_provider: "groq" })
-    // const llm = new ChatAnthropic({ model: "claude-3-" })
-    // const llm = new ChatGroq({ model: "llama3-8b-8192" })
-    // Create agent with the client
-    const agent = new MCPAgent({ llm, client, maxSteps: 30 });
-    // Run the query
-    const result = await agent.run(`Navigate to https://github.com/mcp-use/mcp-use, give a star to the project and write
-a summary of the project.`, 30);
-    console.error(`\nResult: ${result}`);
-}
-if (import.meta.url === `file://${process.argv[1]}`) {
-    main().catch(console.error);
-}

package/dist/examples/chat_example.d.ts

@@ -1,10 +0,0 @@
-/**
- * Simple chat example using MCPAgent with built-in conversation memory.
- *
- * This example demonstrates how to use the MCPAgent with its built-in
- * conversation history capabilities for better contextual interactions.
- *
- * Special thanks to https://github.com/microsoft/playwright-mcp for the server.
- */
-export {};
-//# sourceMappingURL=chat_example.d.ts.map

package/dist/examples/chat_example.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"chat_example.d.ts","sourceRoot":"","sources":["../../examples/chat_example.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG"}

package/dist/examples/chat_example.js

@@ -1,86 +0,0 @@
-/**
- * Simple chat example using MCPAgent with built-in conversation memory.
- *
- * This example demonstrates how to use the MCPAgent with its built-in
- * conversation history capabilities for better contextual interactions.
- *
- * Special thanks to https://github.com/microsoft/playwright-mcp for the server.
- */
-import readline from 'node:readline';
-import { ChatOpenAI } from '@langchain/openai';
-import { config } from 'dotenv';
-import { MCPAgent, MCPClient } from '../index.js';
-// Load environment variables from .env file
-config();
-async function runMemoryChat() {
-    // Config file path - change this to your config file
-    const config = {
-        mcpServers: {
-            airbnb: {
-                command: 'npx',
-                args: ['-y', '@openbnb/mcp-server-airbnb', '--ignore-robots-txt'],
-            },
-        },
-    };
-    console.error('Initializing chat...');
-    // Create MCP client and agent with memory enabled
-    const client = new MCPClient(config);
-    const llm = new ChatOpenAI({ model: 'gpt-4o-mini' });
-    // Create agent with memory_enabled=true
-    const agent = new MCPAgent({
-        llm,
-        client,
-        maxSteps: 15,
-        memoryEnabled: true, // Enable built-in conversation memory
-    });
-    console.error('\n===== Interactive MCP Chat =====');
-    console.error('Type \'exit\' or \'quit\' to end the conversation');
-    console.error('Type \'clear\' to clear conversation history');
-    console.error('==================================\n');
-    // Create readline interface for user input
-    const rl = readline.createInterface({
-        input: process.stdin,
-        output: process.stdout,
-    });
-    const question = (prompt) => {
-        return new Promise((resolve) => {
-            rl.question(prompt, resolve);
-        });
-    };
-    try {
-        // Main chat loop
-        while (true) {
-            // Get user input
-            const userInput = await question('\nYou: ');
-            // Check for exit command
-            if (userInput.toLowerCase() === 'exit' || userInput.toLowerCase() === 'quit') {
-                console.error('Ending conversation...');
-                break;
-            }
-            // Check for clear history command
-            if (userInput.toLowerCase() === 'clear') {
-                agent.clearConversationHistory();
-                console.error('Conversation history cleared.');
-                continue;
-            }
-            // Get response from agent
-            process.stdout.write('\nAssistant: ');
-            try {
-                // Run the agent with the user input (memory handling is automatic)
-                const response = await agent.run(userInput);
-                console.error(response);
-            }
-            catch (error) {
-                console.error(`\nError: ${error}`);
-            }
-        }
-    }
-    finally {
-        // Clean up
-        rl.close();
-        await client.closeAllSessions();
-    }
-}
-if (import.meta.url === `file://${process.argv[1]}`) {
-    runMemoryChat().catch(console.error);
-}

package/dist/examples/filesystem_use.d.ts

@@ -1,11 +0,0 @@
-/**
- * Basic usage example for mcp-use.
- *
- * This example demonstrates how to use the mcp-use library with MCPClient
- * to connect any LLM to MCP tools through a unified interface.
- *
- * Special Thanks to https://github.com/modelcontextprotocol/servers/tree/main/src/filesystem
- * for the server.
- */
-export {};
-//# sourceMappingURL=filesystem_use.d.ts.map

package/dist/examples/filesystem_use.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"filesystem_use.d.ts","sourceRoot":"","sources":["../../examples/filesystem_use.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG"}

package/dist/examples/filesystem_use.js

@@ -1,43 +0,0 @@
-/**
- * Basic usage example for mcp-use.
- *
- * This example demonstrates how to use the mcp-use library with MCPClient
- * to connect any LLM to MCP tools through a unified interface.
- *
- * Special Thanks to https://github.com/modelcontextprotocol/servers/tree/main/src/filesystem
- * for the server.
- */
-import { ChatOpenAI } from '@langchain/openai';
-import { config } from 'dotenv';
-import { MCPAgent, MCPClient } from '../index.js';
-// Load environment variables from .env file
-config();
-const serverConfig = {
-    mcpServers: {
-        filesystem: {
-            command: 'npx',
-            args: [
-                '-y',
-                '@modelcontextprotocol/server-filesystem',
-                'THE_PATH_TO_YOUR_DIRECTORY',
-            ],
-        },
-    },
-};
-async function main() {
-    // Create MCPClient from config
-    const client = MCPClient.fromDict(serverConfig);
-    // Create LLM
-    const llm = new ChatOpenAI({ model: 'gpt-4o' });
-    // const llm = init_chat_model({ model: "llama-3.1-8b-instant", model_provider: "groq" })
-    // const llm = new ChatAnthropic({ model: "claude-3-" })
-    // const llm = new ChatGroq({ model: "llama3-8b-8192" })
-    // Create agent with the client
-    const agent = new MCPAgent({ llm, client, maxSteps: 30 });
-    // Run the query
-    const result = await agent.run('Hello can you give me a list of files and directories in the current directory', 30);
-    console.log(`\nResult: ${result}`);
-}
-if (import.meta.url === `file://${process.argv[1]}`) {
-    main().catch(console.error);
-}

package/dist/examples/http_example.d.ts

@@ -1,18 +0,0 @@
-/**
- * HTTP Example for mcp-use.
- *
- * This example demonstrates how to use the mcp-use library with MCPClient
- * to connect to an MCP server running on a specific HTTP port.
- *
- * Before running this example, you need to start the Playwright MCP server
- * in another terminal with:
- *
- *     npx @playwright/mcp@latest --port 8931
- *
- * This will start the server on port 8931. Resulting in the config you find below.
- * Of course you can run this with any server you want at any URL.
- *
- * Special thanks to https://github.com/microsoft/playwright-mcp for the server.
- */
-export {};
-//# sourceMappingURL=http_example.d.ts.map

package/dist/examples/http_example.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"http_example.d.ts","sourceRoot":"","sources":["../../examples/http_example.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;GAeG"}

package/dist/examples/http_example.js

@@ -1,37 +0,0 @@
-/**
- * HTTP Example for mcp-use.
- *
- * This example demonstrates how to use the mcp-use library with MCPClient
- * to connect to an MCP server running on a specific HTTP port.
- *
- * Before running this example, you need to start the Playwright MCP server
- * in another terminal with:
- *
- *     npx @playwright/mcp@latest --port 8931
- *
- * This will start the server on port 8931. Resulting in the config you find below.
- * Of course you can run this with any server you want at any URL.
- *
- * Special thanks to https://github.com/microsoft/playwright-mcp for the server.
- */
-import { ChatOpenAI } from '@langchain/openai';
-import { config } from 'dotenv';
-import { MCPAgent, MCPClient } from '../index.js';
-// Load environment variables from .env file
-config();
-async function main() {
-    const config = { mcpServers: { http: { url: 'https://gitmcp.io/docs' } } };
-    // Create MCPClient from config
-    const client = MCPClient.fromDict(config);
-    // Create LLM
-    const llm = new ChatOpenAI({ model: 'gpt-4o' });
-    // Create agent with the client
-    const agent = new MCPAgent({ llm, client, maxSteps: 30 });
-    // Run the query
-    const result = await agent.run('Which tools are available and what can they do?', 30);
-    console.log(`\nResult: ${result}`);
-    await agent.close();
-}
-if (import.meta.url === `file://${process.argv[1]}`) {
-    main().catch(console.error);
-}

package/dist/examples/mcp_everything.d.ts

@@ -1,6 +0,0 @@
-/**
- * This example shows how to test the different functionalities of MCPs using the MCP server from
- * anthropic.
- */
-export {};
-//# sourceMappingURL=mcp_everything.d.ts.map

package/dist/examples/mcp_everything.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"mcp_everything.d.ts","sourceRoot":"","sources":["../../examples/mcp_everything.ts"],"names":[],"mappings":"AAAA;;;GAGG"}