mcpserver.so LogoMCPFly
Submit
Spring AI ResOs logo

Spring AI ResOs

by pacphi

AI/IntegrationSpring AIRestaurant BookingAPI-firstChatbotResOs API

This project provides a Spring AI enhanced restaurant booking system using an API-first approach. It includes a code-generated client from the ResOs API, a Spring AI implementation, an MCP server and client, and a ReactJS chatbot UI.

View on GitHub

Last updated: N/A

Developing a Spring AI Enhanced Restaurant Booking System Employing an API-first Approach

GA

GA

Github Action CI Workflow Status

Github Action CI Workflow Status

Known Vulnerabilities

Known Vulnerabilities

This multi-module project hosts a client code-generated from an OpenAPI derivative of the ResOs API combined with a Spring AI implementation. It also includes an MCP server, MCP client configuration for use with Claude and a standalone ReactJS powered chatbot UI.

Background

As a Spring Boot and Spring AI developer, I want to consume libraries that make it convenient to add capabilities to my application(s) as for the following

Use-case:

  • Imagine instead of using OpenTable or Tock you could converse with a chatbot to search for restaurant(s) and make reservation(s) on your behalf.

Getting started

Start with:

  • A Github account
  • (Optional) An API key from ResOS
    • you only need one if you intend to register as a restaurateur!
    • we will spin up a backend that is API-compatible, implemented with Spring Boot Starter Data JDBC
  • An LLM provider
    • e.g., Groq Cloud, OpenRouter, or OpenAI

Prerequisites

  • Git CLI (2.43.0 or better)
  • Github CLI (2.65.0 or better)
  • httpie CLI (3.2.2 or better)
  • Java SDK (21 or better)
  • Maven (3.9.9 or better)
  • an LLM provider account (if using public cloud or commercially hosted models)

How to clone

with Git CLI

git clone https://github.com/pacphi/spring-ai-resos

with Github CLI

gh repo clone pacphi/spring-ai-resos

How to build

Open a terminal shell, then execute:

cd spring-ai-resos
mvn clean install

How to consume

If you want to incorporate any of the starters as dependencies in your own projects, you would:

Add dependency

Maven

<dependency>
    <groupId>me.pacphi</groupId>
    <artifactId>spring-ai-resos-client</artifactId>
    <version>{release-version}</version>
</dependency>

Gradle

implementation 'me.pacphi:spring-ai-resos-client:{release-version}'

Replace occurrences of {release-version} above with a valid artifact release version number

Add configuration

Following Spring Boot conventions, you would add a stanza like this to your:

application.properties

default.url=${RESOS_API_ENDPOINT:https://api.resos.com/v1}

application.yml

default:
  url: ${RESOS_API_ENDPOINT:https://api.resos.com/v1}

To activate the client, specify an API key (if required), and tune other associated configuration.

Consult the chatbot module's configuration for alternative dependencies and configuration that are available to add.

Configuration will be found in labeled spring.config.activate.on-profile sections of the application.yml file.

How to run

You're going to need to launch the backend module first, unless you're a restaurateur, and you have a valid API key for interacting with the ResOS v1.2 API.

To launch the backend, open a terminal shell and execute

cd backend
mvn clean spring-boot:run -Dspring-boot.run.profiles=dev -Dspring-boot.run.jvmArguments="--add-opens java.base/java.net=ALL-UNNAMED"

There's the chatbot module.

But there's also a way to integrate with Claude desktop via MCP client configuration that will consume the MCP server implementation.

with Claude Desktop

Follow these instructions.

Add the following stanza to a file called claude_desktop_config.json:

"spring-ai-resos": {
  "command": "java",
  "args": [
    "-jar",
    "<path-to-project>/target/spring-ai-resos-mcp-server-0.0.1-SNAPSHOT.jar"
  ]
}

or for testing with backend

"spring-ai-resos": {
  "command": "java",
  "args": [
    "-Dspring.profiles.active=dev",
    "-jar",
    "<path-to-project>/target/spring-ai-resos-mcp-server-0.0.1-SNAPSHOT.jar"
  ]
}

Restart Claude Desktop instance. Verify that you have a new set of tool calls available. Chat with Claude.

with Chatbot

Follow these instructions.

To launch the server module, open a terminal shell and execute

cd mcp-server
export RESOS_API_ENDPOINT=http://localhost:8080/api/v1/resos
mvn spring-boot:run -Dspring-boot.run.profiles=cloud,dev

Next, we'll store an API key in a credential file that will allow the chatbot to interact with an LLM service provider.

cd ../mcp-client
leveraging OpenAI

Build and run a version of the chatbot that is compatible for use with OpenAI. You will need to obtain an API key.

Before launching the app:

  • Create a config folder which would be a sibling of the src folder. Create a file named creds.yml inside that folder. Add your own API key into that file.
spring:
  ai:
    openai:
      api-key: {REDACTED}

Replace {REDACTED} above with your OpenAI API key

Next, to launch the chatbot, open a terminal shell and execute

mvn spring-boot:run -Dspring-boot.run.profiles=openai,dev

leveraging Groq Cloud

Build and run a version of the chatbot that is compatible for use with Groq Cloud. You will need to obtain an API key. Note that Groq does not currently have support for text embedding. So if you intend to run with the groq-cloud Spring profile activated, you will also need to provide additional credentials

Before launching the app:

  • Create a config folder which would be a sibling of the src folder. Create a file named creds.yml inside that folder. Add your own API key into that file.
spring:
  ai:
    openai:
      api-key: {REDACTED-1}
      embedding:
        api-key: {REDACTED-2}

Replace {REDACTED-1} and {REDACTED-2} above with your Groq Cloud API and OpenAI keys respectively.

Next, to launch the chatbot, open a terminal shell and execute

mvn spring-boot:run -Dspring-boot.run.profiles=groq-cloud,dev
leveraging OpenRouter

Build and run a version of the chatbot that is compatible for use with OpenRouter. You will need to obtain an API key. Note that OpenRouter does not currently have support for text embedding. So if you intend to run with the openrouter Spring profile activated, you will also need to provide additional credentials

Before launching the app:

  • Create a config folder which would be a sibling of the src folder. Create a file named creds.yml inside that folder. Add your own API key into that file.
spring:
  ai:
    openai:
      api-key: {REDACTED-1}
      embedding:
        api-key: {REDACTED-2}

Replace {REDACTED-1} and {REDACTED-2} above with your OpenRouter API and OpenAI keys respectively.

Next, to launch the chatbot, open a terminal shell and execute

mvn spring-boot:run -Dspring-boot.run.profiles=openrouter,dev

Now, visit http://localhost:8081 in your favorite web-browser.

Spring AI ResOs Chatbot

Spring AI ResOs Chatbot