bedrock-mcp-sample

# Gradio and Amazon Bedrock Integration Project This project is built from scratch to provide a simple front-end interface using Gradio and to perform model inference using the Amazon Bedrock API. ## Project Structure - **README.md**: Project description and documentation. - **pyproject.toml**: Configuration file for the Python project, containing dependency package information used by uv. - **requirements.txt**: List of dependency packages used by traditional pip. - **config/config.json**: Configuration file for Amazon Bedrock. Please specify the Bedrock API endpoint and API key. - **src/frontend.py**: Implementation of the front end using Gradio. It receives user input, queries the model, and displays the results. - **src/bedrock.py**: Responsible for communication with Amazon Bedrock. It reads information from the configuration file and sends API requests. If the endpoint is not set, it returns a simulated response. - **src/mcp_tools/**: Implementation of the Model Context Protocol (MCP) server. Provides tools to fetch content from websites. ## Setup ### Using uv (recommended) 1. Please install uv: ``` pip install uv ``` 2. Please install the dependency packages: ``` uv pip sync pyproject.toml ``` 3. If you also want to install development dependencies: ``` uv pip sync pyproject.toml --extras dev ``` 4. Update `config/config.json` to match your Amazon Bedrock settings: ```json { "bedrock_endpoint": "YOUR_BEDROCK_ENDPOINT", "api_key": "YOUR_API_KEY" } ``` 5. Start the Gradio interface: ``` uv run python src/frontend.py ``` ### Using traditional pip 1. Please install the required packages: ``` pip install -r requirements.txt ``` 2. Update `config/config.json` to match your Amazon Bedrock settings: ```json { "bedrock_endpoint": "YOUR_BEDROCK_ENDPOINT", "api_key": "YOUR_API_KEY" } ``` 3. Start the Gradio interface: ``` python src/frontend.py ``` ## Usage Please enter your query in the text box displayed in the Gradio UI. The input will be sent to Amazon Bedrock via `src/bedrock.py`, and the results will be displayed on the screen. ## Notes - If the `bedrock_endpoint` is not set, a simulated response will be returned. - Please ensure that the API credentials and network settings are correctly configured. Feel free to use it.