agentmakemcp 0.0.1__py3-none-any.whl

agentmakemcp/README.md

@@ -0,0 +1,143 @@
+=========================================================
+AgentMake MCP: Your Gateway to Multi-Agent AI Systems
+=========================================================
+
+**agentmake_mcp** offers the simplest way to set up Modal Context Protocol (MCP) servers, powering them with the versatile agentic components from the `AgentMake AI <https://github.com/eliranwong/agentmake>`_ framework. This project provides the essential tools and infrastructure to create sophisticated multi-agent systems that can tackle complex tasks through collaboration and dynamic task allocation.
+
+While **AgentMake AI** provides the core building blocks for creating individual AI agents, **agentmake_mcp** enables you to assemble and orchestrate them. Think of **AgentMake AI** as the factory for creating your specialized AI workers, and **agentmake_mcp** as the central command center where you manage your teams of agents on large-scale projects.
+
+With AgentMake MCP, you can:
+
+*   **Integrate Diverse AI Tools:** Don't limit yourself to a single AI tool. Seamlessly connect AgentMake AI components with third-party AI tools using the standardized MCP interface.
+*   **Orchestrate Multiple Agents:** Define and manage how different AI agents, each with unique skills, collaborate to achieve a common goal.
+*   **Build Complex Workflows:** Design intricate workflows where agents delegate tasks, share information, and work in parallel to solve problems more efficiently.
+*   **Host Your Own MCP Server:** Easily deploy a dedicated server for your multi-agent systems, allowing for robust integration with your applications and services.
+
+Key Features
+============
+
+*   **Easy Server Setup:** Quickly deploy a fully functional MCP server with minimal configuration.
+*   **Seamless AgentMake AI Integration:** Leverage the full power of AgentMake AI's 16+ AI backends and 7 agentic components.
+*   **MCP Standards:** Supports running both MCP tools and MCP prompts with `AgentMake` components.
+*   **Flexible Agent Orchestration:** Define custom collaboration strategies and communication protocols for your agents.
+*   **Scalable and Extensible:** Designed to support a growing number of agents and complex workflows.
+*   **Developer-Friendly:** A clean and intuitive API for defining and managing your multi-agent systems.
+
+Getting Started
+===============
+
+1. Prerequisites
+----------------
+
+*   Python 3.8+
+*   Familiarity with the agentic components supported by `AgentMake AI <https://github.com/eliranwong/agentmake>`_.
+
+`agentmake_mcp` automatically includes the `agentmake` library.
+
+2. Installation
+---------------
+
+.. code-block:: bash
+
+    pip install --upgrade agentmake_mcp
+
+To include support for Google's Vertex AI, install with the `[genai]` extra:
+
+.. code-block:: bash
+
+    pip install --upgrade agentmake_mcp[genai]
+
+3. Create a Configuration File
+------------------------------
+
+Create a Python file (e.g., `my_mcp_server.py`) and define a dictionary that configures your server. This dictionary can be named anything, as the server will automatically discover it.
+
+Here is the structure of the configuration dictionary:
+
+*   `server` (required): `str` - The name of your MCP server.
+*   `transport` (optional): `str` - The transport protocol. Defaults to `http`.
+*   `port` (optional): `int` - The server port. Defaults to `8080`.
+*   `settings` (required): `list[dict]` - A list of dictionaries, where each dictionary defines an MCP prompt or tool.
+
+Each dictionary placed in the settings list may have the following keys and values:
+
+* name [required/optional] - This field is mandatory unless a tool is specified. It is a string value that serves as an identifier for a MCP prompt or tool.
+* description [required/optional] - This field is also mandatory unless a tool is specified. It is a string that provides a detailed description of a MCP prompt or tool.
+* agentmake [required] - This is a required field that can be either a string or a dictionary. To add a MCP prompt, a string value should be provided. Alternatively, to add a MCP tool, a dictionary should be used.
+
+Setting up an MCP Prompt
+~~~~~~~~~~~~~~~~~~~~~~~~
+
+To add an MCP prompt, provide the prompt string directly as the value for the `agentmake` key.
+
+Setting up an MCP Tool
+~~~~~~~~~~~~~~~~~~~~~~
+
+To add an MCP tool, provide a dictionary for the `agentmake` key. This dictionary specifies the parameters for the `agentmake` signature function from the AgentMake AI library (excluding the `messages` parameter).
+
+*For more details on the `agentmake` function parameters, see the `AgentMake AI documentation <https://github.com/eliranwong/agentmake/blob/main/docs/README.md>`_.*
+
+4. Running the Server
+---------------------
+
+Run the `agentmakemcp` command from your terminal, passing your configuration file as an argument.
+
+**For examples**
+
+.. code-block:: bash
+
+    agentmakemcp examples/ask_multiple_models.py
+
+.. code-block:: bash
+
+    agentmakemcp examples/different_persona.py
+
+.. code-block:: bash
+
+    agentmakemcp examples/youtube_utilities.py
+
+.. code-block:: bash
+
+    agentmakemcp teamwork_and_toolmate.py
+
+**Remarks:**
+
+*   You can run multiple AgentMake MCP servers simultaneously on different ports.
+*   You can specify different AI backends for different tools, even on the same MCP server.
+
+More Examples
+=============
+
+You can find more advanced examples, such as chaining multiple agents together, in the `/examples <https://github.com/eliranwong/agentmake_mcp/tree/main/examples>`_ directory.
+
+Integration with Third-Party AI Tools:
+========================================
+
+For example, to integrate `AgentMake MCP servers` with `Gemini CLI`:
+
+.. code-block:: bash
+
+    agentmakemcp examples/ask_multiple_models.py
+
+Edit `.gemini/settings.json` to include the following block:
+
+.. code-block:: json
+
+    {
+      // add MCP servers
+      "mcpServers": {
+        "Ask Multiple AI Models": {
+          "httpUrl": "http://127.0.0.1:8080/mcp/"
+        }
+      }
+    }
+
+Contributing
+============
+
+We welcome contributions from the community! If you have an idea for a new feature, a bug fix, or an improvement to the documentation, please open an issue or submit a pull request.
+
+License
+=======
+
+This project is licensed under the MIT License. See the `LICENSE file <https://github.com/eliranwong/agentmake_mcp/blob/main/LICENSE>`_ for details.

agentmakemcp/__init__.py

@@ -0,0 +1,157 @@
+from agentmake import DEVELOPER_MODE, USER_OS, AGENTMAKE_USER_DIR, PACKAGE_PATH, readTextFile, writeTextFile
+from pathlib import Path
+import os, traceback, re
+
+
+def readToolFile(tool_object: str) -> set:
+    """return tool description and parameters"""
+    tool_name = tool_object[:20]
+    if USER_OS == "Windows":
+        tool_object = os.path.join(*tool_object.split("/"))
+    possible_tool_file_path_2 = os.path.join(PACKAGE_PATH, "tools", f"{tool_object}.py")
+    possible_tool_file_path_1 = os.path.join(AGENTMAKE_USER_DIR, "tools", f"{tool_object}.py")
+    if tool_object is None:
+        pass
+    elif os.path.isfile(possible_tool_file_path_1):
+        tool_file_content = readTextFile(possible_tool_file_path_1)
+        if tool_file_content:
+            tool_object = tool_file_content
+    elif os.path.isfile(possible_tool_file_path_2):
+        tool_file_content = readTextFile(possible_tool_file_path_2)
+        if tool_file_content:
+            tool_object = tool_file_content
+    elif os.path.isfile(tool_object): # tool_object itself is a valid filepath
+        tool_file_content = readTextFile(tool_object)
+        if tool_file_content:
+            tool_object = tool_file_content
+    if tool_object:
+        glob = {}
+        loc = {}
+        try:
+            exec(tool_object, glob, loc)
+            schema = loc.get("TOOL_SCHEMA")
+            description = description = schema["description"] if schema else loc.get("TOOL_DESCRIPTION", "")
+            parameters = schema["parameters"] if schema and "parameters" in schema else {}
+            return description, parameters
+        except Exception as e:
+            print(f"Failed to execute tool `{tool_name}`! An error occurred: {e}")
+            if DEVELOPER_MODE:
+                print(traceback.format_exc())
+    return "", {}
+
+def compileServerScripts(configs: dict) -> str:
+    scripts = []
+    ob, cb = "{", "}"
+    server = configs.get("server", "AgentMake AI")
+    port = configs.get("port", 8080)
+    transport = configs.get("transport", "http")
+
+    for i in configs["settings"]:
+        if "agentmake" in i:
+            am = i["agentmake"]
+
+            if isinstance(am, str): # add a mcp prompt
+                # add to server script
+                script = f"""@mcp.tool
+def {i["name"].replace(" ", "_").lower()}(request:str) -> PromptMessage:
+    \"\"\"{i["description"]}\"\"\"
+    logger.info(f"[Request] {ob}request{cb}")
+    prompt_text = \"\"\"{am}
+
+# Here is the request:
+---
+{ob}request{cb}
+---
+\"\"\"
+    return PromptMessage(role="user", content=TextContent(type="text", text=prompt_text))"""
+                scripts.append(script)
+
+            elif isinstance(am, dict): # add a mcp tool
+                # fill in tool name and description when a tool is specified
+                if not "agent" in am and "tool" in am and am["tool"]:
+                    firstTool = am["tool"] if isinstance(am["tool"], str) else am["tool"][0]
+                    if not "name" in i:
+                        i["name"] = firstTool.replace("/", "_")
+                    description, parameters = readToolFile(firstTool)
+                    if not "description" in i:
+                        i["description"] = description
+                    # specify tool arguments in tool description
+                    if parameters and not re.search(r"^Args \[", description):
+                        required = parameters["required"]
+                        optional = [o for o in parameters["properties"] if not o in required]
+                        if required:
+                            args = [f"{r}: {parameters["properties"][r]['description']}" for r in required]
+                            args = "\n    ".join(args)
+                            i["description"] += f"""
+
+Args [required]:
+    {args}
+"""
+                        if optional:
+                            args = [f"{o}: {parameters["properties"][o]['description']}" for o in optional]
+                            args = "\n    ".join(args)
+                            i["description"] += f"""
+Args [optional]:
+    {args}
+"""
+
+                # check name and description
+                if not "name" in i:
+                    print(f"Error! Parameter `name` is missing in item {i}!")
+                    exit(1)
+                elif not "description" in i:
+                    print(f"Error! Parameter `description` is missing in item {i}!")
+                    exit(1)
+                
+                # add to server script
+                script = f"""@mcp.tool
+def {i["name"].replace(" ", "_").lower()}(request:str) -> str:
+    \"\"\"{i["description"]}\"\"\"
+    logger.info(f"[Request] {ob}request{cb}")
+    messages = agentmake(request, **{am})
+    return getResponse(messages)"""
+                scripts.append(script)
+
+    mcp_server_script = f"""from fastmcp import FastMCP
+from fastmcp.prompts.prompt import Message, PromptMessage, TextContent
+from agentmake import agentmake
+import logging, os
+
+logger = logging.getLogger(__name__)
+logging.basicConfig(format="[%(levelname)s]: %(message)s", level=logging.INFO)
+
+mcp = FastMCP(name="{server}")
+
+def getResponse(messages:list) -> str:
+    return messages[-1].get("content") if messages and "content" in messages[-1] else "Error!"
+
+""" + "\n\n".join(scripts) + f"""
+
+logger.info(f"MCP server `{server}` started on port {port}")
+mcp.run(transport="{transport}", port={port})"""
+
+    # back up mcp server script
+    folderPath = os.path.join(AGENTMAKE_USER_DIR, "mcp")
+    Path(folderPath).mkdir(parents=True, exist_ok=True)
+    filePath = os.path.join(folderPath, f"{server.replace(' ', '_')}.py")
+    writeTextFile(filePath, mcp_server_script)
+    print(f"\n# Backup\n\nAgentMake MCP Server script was generated and saved to: {filePath}\n")
+
+    # setup hints
+    print(f"""# Starting MCP server `{server}` on port `{port}` ...
+
+# Integration with Other AI Tools
+          
+For example, to work with `Gemini CLI`, add the following block: in `~/.gemini/settings.json`:
+          
+{ob}
+  ...
+  "mcpServers": {ob}
+    "{server}": {ob}
+      "httpUrl": "http://127.0.0.1:{port}/mcp/"
+    {cb}
+  {cb}
+{cb}
+""")
+
+    return mcp_server_script

agentmakemcp/main.py

@@ -0,0 +1,30 @@
+from agentmake import readTextFile
+from agentmakemcp import compileServerScripts
+import os, argparse
+
+def main():
+    parser = argparse.ArgumentParser()
+    parser.add_argument("default", nargs="*", default=None, help="AgentMake AI MCP server configuration file path.")
+    args = parser.parse_args()
+
+    if args.default is None:
+        print("Error! You need to specify an AgentMake AI MCP server configuration file path, e.g. `examples/youtube_utilities.py`")
+        exit(1)
+
+    filepath = args.default
+    if os.path.isfile(filepath):
+        configs = readTextFile(filepath)
+        if configs:
+            configs = eval(configs)
+            if isinstance(configs, dict):
+                exec(compileServerScripts(configs))
+            else:
+                print("Error! Invalid config file! Read instructions at: https://github.com/eliranwong/agentmake_mcp")
+                exit(1)
+    else:
+        print(f"Error! File `{filepath}` does not exist!")
+        exit(1)
+
+
+if __name__ == "__main__":
+    main()

agentmakemcp/package_name.txt

@@ -0,0 +1 @@
+agentmakemcp

agentmakemcp/requirements.txt

@@ -0,0 +1 @@
+agentmake

agentmakemcp-0.0.1.dist-info/METADATA

@@ -0,0 +1,172 @@
+Metadata-Version: 2.1
+Name: agentmakemcp
+Version: 0.0.1
+Summary: AgentMake AI MCP Servers - Easy setup of MCP servers running AgentMake AI agentic components.
+Home-page: https://toolmate.ai
+Author: Eliran Wong
+Author-email: support@toolmate.ai
+License: GNU General Public License (GPL)
+Project-URL: Source, https://github.com/eliranwong/agentmake_mcp
+Project-URL: Tracker, https://github.com/eliranwong/agentmake_mcp/issues
+Project-URL: Documentation, https://github.com/eliranwong/agentmake_mcp/wiki
+Project-URL: Funding, https://www.paypal.me/toolmate
+Keywords: mcp agent toolmate ai anthropic azure chatgpt cohere deepseek genai github googleai groq llamacpp mistral ollama openai vertexai xai
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: End Users/Desktop
+Classifier: Topic :: Utilities
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Topic :: Software Development :: Build Tools
+Classifier: License :: OSI Approved :: GNU General Public License v3 or later (GPLv3+)
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Requires-Python: >=3.8, <3.13
+Requires-Dist: agentmake
+Provides-Extra: genai
+Requires-Dist: google-genai >=1.25.0 ; extra == 'genai'
+
+=========================================================
+AgentMake MCP: Your Gateway to Multi-Agent AI Systems
+=========================================================
+
+**agentmake_mcp** offers the simplest way to set up Modal Context Protocol (MCP) servers, powering them with the versatile agentic components from the `AgentMake AI <https://github.com/eliranwong/agentmake>`_ framework. This project provides the essential tools and infrastructure to create sophisticated multi-agent systems that can tackle complex tasks through collaboration and dynamic task allocation.
+
+While **AgentMake AI** provides the core building blocks for creating individual AI agents, **agentmake_mcp** enables you to assemble and orchestrate them. Think of **AgentMake AI** as the factory for creating your specialized AI workers, and **agentmake_mcp** as the central command center where you manage your teams of agents on large-scale projects.
+
+With AgentMake MCP, you can:
+
+*   **Integrate Diverse AI Tools:** Don't limit yourself to a single AI tool. Seamlessly connect AgentMake AI components with third-party AI tools using the standardized MCP interface.
+*   **Orchestrate Multiple Agents:** Define and manage how different AI agents, each with unique skills, collaborate to achieve a common goal.
+*   **Build Complex Workflows:** Design intricate workflows where agents delegate tasks, share information, and work in parallel to solve problems more efficiently.
+*   **Host Your Own MCP Server:** Easily deploy a dedicated server for your multi-agent systems, allowing for robust integration with your applications and services.
+
+Key Features
+============
+
+*   **Easy Server Setup:** Quickly deploy a fully functional MCP server with minimal configuration.
+*   **Seamless AgentMake AI Integration:** Leverage the full power of AgentMake AI's 16+ AI backends and 7 agentic components.
+*   **MCP Standards:** Supports running both MCP tools and MCP prompts with `AgentMake` components.
+*   **Flexible Agent Orchestration:** Define custom collaboration strategies and communication protocols for your agents.
+*   **Scalable and Extensible:** Designed to support a growing number of agents and complex workflows.
+*   **Developer-Friendly:** A clean and intuitive API for defining and managing your multi-agent systems.
+
+Getting Started
+===============
+
+1. Prerequisites
+----------------
+
+*   Python 3.8+
+*   Familiarity with the agentic components supported by `AgentMake AI <https://github.com/eliranwong/agentmake>`_.
+
+`agentmake_mcp` automatically includes the `agentmake` library.
+
+2. Installation
+---------------
+
+.. code-block:: bash
+
+    pip install --upgrade agentmake_mcp
+
+To include support for Google's Vertex AI, install with the `[genai]` extra:
+
+.. code-block:: bash
+
+    pip install --upgrade agentmake_mcp[genai]
+
+3. Create a Configuration File
+------------------------------
+
+Create a Python file (e.g., `my_mcp_server.py`) and define a dictionary that configures your server. This dictionary can be named anything, as the server will automatically discover it.
+
+Here is the structure of the configuration dictionary:
+
+*   `server` (required): `str` - The name of your MCP server.
+*   `transport` (optional): `str` - The transport protocol. Defaults to `http`.
+*   `port` (optional): `int` - The server port. Defaults to `8080`.
+*   `settings` (required): `list[dict]` - A list of dictionaries, where each dictionary defines an MCP prompt or tool.
+
+Each dictionary placed in the settings list may have the following keys and values:
+
+* name [required/optional] - This field is mandatory unless a tool is specified. It is a string value that serves as an identifier for a MCP prompt or tool.
+* description [required/optional] - This field is also mandatory unless a tool is specified. It is a string that provides a detailed description of a MCP prompt or tool.
+* agentmake [required] - This is a required field that can be either a string or a dictionary. To add a MCP prompt, a string value should be provided. Alternatively, to add a MCP tool, a dictionary should be used.
+
+Setting up an MCP Prompt
+~~~~~~~~~~~~~~~~~~~~~~~~
+
+To add an MCP prompt, provide the prompt string directly as the value for the `agentmake` key.
+
+Setting up an MCP Tool
+~~~~~~~~~~~~~~~~~~~~~~
+
+To add an MCP tool, provide a dictionary for the `agentmake` key. This dictionary specifies the parameters for the `agentmake` signature function from the AgentMake AI library (excluding the `messages` parameter).
+
+*For more details on the `agentmake` function parameters, see the `AgentMake AI documentation <https://github.com/eliranwong/agentmake/blob/main/docs/README.md>`_.*
+
+4. Running the Server
+---------------------
+
+Run the `agentmakemcp` command from your terminal, passing your configuration file as an argument.
+
+**For examples**
+
+.. code-block:: bash
+
+    agentmakemcp examples/ask_multiple_models.py
+
+.. code-block:: bash
+
+    agentmakemcp examples/different_persona.py
+
+.. code-block:: bash
+
+    agentmakemcp examples/youtube_utilities.py
+
+.. code-block:: bash
+
+    agentmakemcp teamwork_and_toolmate.py
+
+**Remarks:**
+
+*   You can run multiple AgentMake MCP servers simultaneously on different ports.
+*   You can specify different AI backends for different tools, even on the same MCP server.
+
+More Examples
+=============
+
+You can find more advanced examples, such as chaining multiple agents together, in the `/examples <https://github.com/eliranwong/agentmake_mcp/tree/main/examples>`_ directory.
+
+Integration with Third-Party AI Tools:
+========================================
+
+For example, to integrate `AgentMake MCP servers` with `Gemini CLI`:
+
+.. code-block:: bash
+
+    agentmakemcp examples/ask_multiple_models.py
+
+Edit `.gemini/settings.json` to include the following block:
+
+.. code-block:: json
+
+    {
+      // add MCP servers
+      "mcpServers": {
+        "Ask Multiple AI Models": {
+          "httpUrl": "http://127.0.0.1:8080/mcp/"
+        }
+      }
+    }
+
+Contributing
+============
+
+We welcome contributions from the community! If you have an idea for a new feature, a bug fix, or an improvement to the documentation, please open an issue or submit a pull request.
+
+License
+=======
+
+This project is licensed under the MIT License. See the `LICENSE file <https://github.com/eliranwong/agentmake_mcp/blob/main/LICENSE>`_ for details.

agentmakemcp-0.0.1.dist-info/RECORD

@@ -0,0 +1,10 @@
+agentmakemcp/README.md,sha256=vs8U4aSOxDHGBWfAu_rEpMjzCrThMCbZ_8eHCNo0Ke4,6339
+agentmakemcp/__init__.py,sha256=IDXupJngUjkTMI8Sge-Lh8R2x0QygJ28MJHMWsG96dc,6136
+agentmakemcp/main.py,sha256=PIlNJ-dgHiHHHOGTBHofjY46PRTNRNseiwOu9p-j2dw,1018
+agentmakemcp/package_name.txt,sha256=fnDPfL914e8QqQoYRKa1HlvfoYE1jaP7DmG_58zR8zg,12
+agentmakemcp/requirements.txt,sha256=MoJ3-djAIb8rlOJWSdXciyo2AMYJBq1xXQaI4MZ-N9Q,9
+agentmakemcp-0.0.1.dist-info/METADATA,sha256=a3IMqKF_NavpMf16IHyMaazeRAbOXgIPBVyO2c1-Kfk,7776
+agentmakemcp-0.0.1.dist-info/WHEEL,sha256=oiQVh_5PnQM0E3gPdiz09WCNmwiHDMaGer_elqB3coM,92
+agentmakemcp-0.0.1.dist-info/entry_points.txt,sha256=qHK0Au73BPK4DhYYy86XlQz2MxUTqCiU6FjjXhDZtF4,56
+agentmakemcp-0.0.1.dist-info/top_level.txt,sha256=fFBZIL_ORW_naYA46Q3eTsNhZVwk2pl4uaF1YiSV_4s,13
+agentmakemcp-0.0.1.dist-info/RECORD,,

agentmakemcp-0.0.1.dist-info/WHEEL

@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: bdist_wheel (0.42.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+

agentmakemcp-0.0.1.dist-info/entry_points.txt

@@ -0,0 +1,2 @@
+[console_scripts]
+agentmakemcp = agentmakemcp.main:main

agentmakemcp-0.0.1.dist-info/top_level.txt

@@ -0,0 +1 @@
+agentmakemcp