Merge pull request #16 from hideya/dev

Update to ver 0.3.0

Author: hideya

.env.template

@@ -1,8 +1,10 @@
-
+# https://console.anthropic.com/settings/keys
 ANTHROPIC_API_KEY=sk-ant-...
+# https://platform.openai.com/api-keys
 OPENAI_API_KEY=sk-proj-...
+# https://aistudio.google.com/apikey
 GOOGLE_API_KEY=AI...
-# GROQ_API_KEY=gsk_...
 
-# GITHUB_PERSONAL_ACCESS_TOKEN=github_pat_...
 # BRAVE_API_KEY=BSA...
+# GITHUB_PERSONAL_ACCESS_TOKEN=github_pat_...
+# NOTION_INTEGRATION_SECRET=ntn_...

README.md

@@ -4,10 +4,17 @@
 **Quickly test and explore MCP servers from the command line!**
 
 A simple, text-based CLI client for [Model Context Protocol (MCP)](https://modelcontextprotocol.io/) servers built with LangChain and TypeScript.  
+This tool performs automatic schema adjustments for LLM compatibility.  
 Suitable for testing MCP servers, exploring their capabilities, and prototyping integrations.
 
 Internally it uses  [LangChain ReAct Agent](https://github.com/langchain-ai/react-agent-js) and
-a utility function `convertMcpToLangchainTools()` from [`@h1deya/langchain-mcp-tools`](https://www.npmjs.com/package/@h1deya/langchain-mcp-tools).
+a utility function `convertMcpToLangchainTools()` from
+[`@h1deya/langchain-mcp-tools`](https://www.npmjs.com/package/@h1deya/langchain-mcp-tools).  
+This function performs the aforementioned MCP tools schema transformations for LLM compatibility.
+See [this page](https://github.com/hideya/langchain-mcp-tools-ts/blob/main/README.md#llm-provider-schema-compatibility)
+for details.
+
+A Python equivalent of this utility is available [here](https://pypi.org/project/mcp-chat/)
 
 ## Prerequisites
 
@@ -18,7 +25,7 @@ a utility function `convertMcpToLangchainTools()` from [`@h1deya/langchain-mcp-t
   [OpenAI](https://platform.openai.com/api-keys),
   [Anthropic](https://console.anthropic.com/settings/keys),
   and/or
-  [Google GenAI](https://aistudio.google.com/apikey)
+  [Google AI Studio (for GenAI/Gemini)](https://aistudio.google.com/apikey)
   as needed
 
 ## Quick Start
@@ -43,7 +50,7 @@ a utility function `convertMcpToLangchainTools()` from [`@h1deya/langchain-mcp-t
       // "model_provider": "anthropic",
       // "model": "claude-3-5-haiku-latest",
       // "model_provider": "google_genai",
-      // "model": "gemini-2.0-flash",
+      // "model": "gemini-2.5-flash",
     },
 
     "mcp_servers": {
@@ -128,9 +135,9 @@ mcp-try-cli --help
 
 ## Supported LLM Providers
 
-- **OpenAI**: `gpt-4o`, `gpt-4o-mini`, etc.
+- **OpenAI**: `o4-mini`, `gpt-4o-mini`, etc.
 - **Anthropic**: `claude-sonnet-4-0`, `claude-3-5-haiku-latest`, etc.
-- **Google (GenAI)**: `gemini-2.0-flash`, `gemini-1.5-pro`, etc.
+- **Google (GenAI)**: `gemini-2.5-pro`, `gemini-2.5-flash`, etc.
 
 ## Configuration
 
@@ -153,7 +160,7 @@ Create a `llm_mcp_config.json5` file:
 {
   "llm": {
     "model_provider": "openai",
-    "model": "gpt-4o-mini",
+    "model": "gpt-4.1-nano",
     // model: "o4-mini",
   },
 
@@ -165,8 +172,8 @@ Create a `llm_mcp_config.json5` file:
 
   // "llm": {
   //   "model_provider": "google_genai",
-  //   "model": "gemini-2.0-flash",
-  //   // "model": "gemini-2.5-pro-preview-06-05",
+  //   "model": "gemini-2.5-flash",
+  //   // "model": "gemini-2.5-pro",
   // }
 
   "example_queries": [
@@ -246,14 +253,6 @@ There are quite a few useful MCP servers already available:
 - Use `--verbose` flag for detailed output
 - Refer to [MCP documentation](https://modelcontextprotocol.io/)
 
-## Development
-
-This tool is built with:
-- [Model Context Protocol (MCP)](https://modelcontextprotocol.io/)
-- [LangChain](https://langchain.com/) for LLM integration
-- [TypeScript](https://www.typescriptlang.org/) for type safety
-- [Yargs](https://yargs.js.org/) for CLI parsing
-
 ## License
 
 MIT License - see [LICENSE](LICENSE) file for details.

README_DEV.md

@@ -60,7 +60,7 @@ For the convenience of debugging MCP servers, this client prints local (stdio) M
 
 LLMs from Anthropic, OpenAI and Google (GenAI) are currently supported.
 
-A python version of this MCP client is available
+A Python version of this MCP client is available
 [here](https://github.com/hideya/mcp-client-langchain-py)
 
 ## Prerequisites

llm_mcp_config.json5

@@ -15,32 +15,33 @@
     //     // "max_tokens": 10000,
     // },
 
-    "llm": {
-        // https://platform.openai.com/docs/pricing
-        // https://platform.openai.com/settings/organization/billing/overview
-        "model_provider": "openai",
-        "model": "gpt-4o-mini",
-        // "model": "o4-mini",
-        // "temperature": 0.0,  //  'temperature' is not supported with "o4-mini"
-        // "max_completion_tokens": 10000,  //  Use 'max_completion_tokens' instead of 'max_tokens'
-    },
-
     // "llm": {
-    //     // https://ai.google.dev/gemini-api/docs/pricing
-    //     // https://console.cloud.google.com/billing
-    //     "model_provider": "google_genai",
-    //     "model": "gemini-2.0-flash",
-    //     // "model": "gemini-1.5-pro",
-    //     // "temperature": 0.0,
-    //     // "max_tokens": 10000,
+    //     // https://platform.openai.com/docs/pricing
+    //     // https://platform.openai.com/settings/organization/billing/overview
+    //     "model_provider": "openai",
+    //     "model": "gpt-4.1-nano",
+    //     // "model": "o4-mini",
+    //     // "temperature": 0.0,  //  'temperature' is not supported with "o4-mini"
+    //     // "max_completion_tokens": 10000,  //  Use 'max_completion_tokens' instead of 'max_tokens'
     // },
 
+    "llm": {
+        // https://ai.google.dev/gemini-api/docs/pricing
+        // https://console.cloud.google.com/billing
+        "model_provider": "google_genai",
+        "model": "gemini-2.5-flash",
+        // "model": "gemini-2.5-pro",
+        // "temperature": 0.0,
+        // "max_tokens": 10000,
+    },
+
     "example_queries": [
-        "Are there any weather alerts in California?",
         "Read the news headlines on bbc.com",
         "Read and briefly summarize the LICENSE file",
+        "Are there any weather alerts in California?",
+        // "What's the news from Tokyo today?",
         // "Open the webpage at bbc.com",
-        // "Search the web and get today's news related to tokyo",
+        // "Tell me about my Notion account",
     ],
 
     "mcp_servers": {
@@ -82,7 +83,7 @@
         // },
 
         // // Test SSE connection with the auto fallback
-        // // See the comments at the top of index.ts
+        // // See the comments at the top of src/index.ts
         // weather: {
         //     "url": "http://localhost:${SSE_SERVER_PORT}/sse"
         // },
@@ -112,5 +113,16 @@
         //     "args": [ "-y", "@modelcontextprotocol/server-brave-search"],
         //     "env": { "BRAVE_API_KEY": "${BRAVE_API_KEY}" }
         // },
+
+        // notion: {
+        //   "command": "npx",
+        //   "args": ["-y", "@notionhq/notion-mcp-server"],
+        //   "env": {
+        //     // Although the following implies that this MCP server is designed for
+        //     // OpenAI LLMs, it works fine with others models.
+        //     // Tested Claude and Gemini (with schema adjustments).
+        //     "OPENAPI_MCP_HEADERS": '{"Authorization": "Bearer ${NOTION_INTEGRATION_SECRET}", "Notion-Version": "2022-06-28"}'
+        //   },
+        // },
     }
 }