## Agent and MCP Deployment Host Service - DeepNLP x aiagenta2z.com [Website](https://www.deepnlp.org/workspace/deploy) | [GitHub](https://github.com/aiagenta2z/agent_mcp_deployment)| [AI Agent Marketplace](https://www.deepnlp.org/store/ai-agent) | [AI Agent A2Z](https://www.aiagenta2z.com) DeepNLP x AI Agent A2Z provide public hosting service of AI Agent and MCP Deployment. Users can get a unique live subdomain endpoint for their project, which can be used for ChatGPT, Cursor to connect and use your agents and MCPs, e.g. Live URL: `${owner_name}.aiagenta2z.com/${repo_name}/mcp` [Visit Deployment Panel](https://www.deepnlp.org/workspace/deploy) ### **Features** 1. Various Deployment methods: template, github_repo, and deploy from source code 2. Templates: We provides 20+ templates in business models, such as `selling product` and `digital resources` e-commerce products agent/mcps as resources, verdors and content creators can expose their physical goods, digital resources (documents,files,online courses) etc to ChatGPT/Cursor. 3. GitHub/Source Code: Support both Python/Typescript, which is just like how you start your Agent locally, you can deploy in our container without renting the cloud server yourself. 4. Subdomain URL: Each User will have a unique subdomain URL for your agents, able to verification and hosting services. ## Quickstart ### Deploy From Template (Beta) #### Use Case 1 Selling products ### Deploy From Source Code #### User Case 2 Hosting a Google Deep Research Agent MCP <img src="https://raw.githubusercontent.com/aiagenta2z/agent_mcp_deployment/refs/heads/main/docs/agent_mcp_deployment_panel.png" style="height:400px;" alt="AI Agent Marketplace Category"> Lets' say you want to implement a Google Customized Search API based MCP server and wants to expose a tool `google_search(query: str, num: int = 10, full_response: bool = False)` for users to use. And you have already prepared below information: Requirements ```angular2html ## Create a new project and Register AI Service unique_id: derek/google-deep-search-agent ## archived source code google-deep-search-agent.zip ## Starting Command uvicorn server:app ## pip requirement files in the archived source code google-deep-search-agent.zip requirements.txt ``` Step 1. Goto `Workspace->Agent Deployment` and visit the [Deployment Workspace](https://www.deepnlp.org/workspace/deploy) Select the project to deploy 'derek/google-deep-search-agent' Step 2. Switch Tab: Custom Python/JS Drag and drop the source code archive file `google-deep-search-agent.zip` to upload Step 3. Choose Config and Deploy `Deploy Region`: `global` for avoid most ip restriction) `Entry Point`: Input `uvicorn server:app`, this is the command that you use to start MCP/Agent server locally, for example we have a server.py file and an app class, and we use `uvicorn` to start the mcp, please avoid specifying any `ports` and we use assign dynamically. `Environment Variables`: We put `GOOGLE_SEARCH_ACCESS_KEY` and the key value in this field. Note that your access key is safe and we will use pass the keys as variables in the requests to start your service in the container. It's equivalent to `.env` files in your uploaded sources. Step 4. Deploy Click deploy button and please please wait a while for the deployment to complete and you will find your subdomain live url ready! MCP SERVER URL: `derekzz.aiagenta2z.com/google-deep-search-agent` ### Deploy From GitHub Repo (Beta) Step 1. Switch Tab: GitHub Input your source code to git clone and deploy on the server. e.g. https://github.com/org/repo Step 2. Choose Config and Deploy `Deploy Region`: `global` for avoid most ip restriction) `Entry Point`: Input `uvicorn server:app`, this is the command that you use to start MCP/Agent server locally, for example we have a server.py file and an app class, and we use `uvicorn` to start the mcp, please avoid specifying any `ports` and we use assign dynamically. `Environment Variables`: We put `GOOGLE_SEARCH_ACCESS_KEY` and the key value in this field. Note that your access key is safe and we will use pass the keys as variables in the requests to start your service in the container. It's equivalent to `.env` files in your uploaded sources. Step 3. Deploy Click deploy button and please wait a while for the deployment process to complete and you will find your subdomain live url ready! ### Related [AI Agent Marketplace Registry](https://github.com/aiagenta2z/ai-agent-marketplace) [Open AI Agent Marketplace](https://www.deepnlp.org/store/ai-agent) [MCP Marketplace](https://www.deepnlp.org/store/ai-agent/mcp-server) [OneKey Router AI Agent & MCP Ranking](https://www.deepnlp.org/agent/rankings) [OneKey Agent MCP Router](https://www.deepnlp.org/agent/onekey-mcp-router) [OneKey AGent MCP Router Doc](https://deepnlp.org/doc/onekey_mcp_router) [AI Agent Dataset](https://www.deepnlp.org/store/dataset) [Gemini Nano Banana Agent](https://agent.deepnlp.org/agent/mcp_tool_use?server=aiagenta2z%2Fgemini_mcp_onekey)