Multi-Agent AI Meal Planner

This application uses LangChain, Anthropic's Claude, Kafka, and Flink to create a meal plan and grocery list for a family with young children.

This project demonstrates how to use an event-driven architecture to coordinate a team of agents to solve a problem. In this model, agents are designed to emit and listen for events. Events are signals that something has happened, to which agents can respond without otherwise orchestrating a direct request to act.

How it works

As a user, you can configure settings for your family such as your children's likes and dislikes and how many meals per week you want to generate.

Clicking on Request a Meal Plan writes a record to a backend database served by MongoDB. Once a record is written to the database, an event with the record information is written to a Kafka topic on Confluent Cloud.

This starts the multi-agent process. There's currently four different agents working behind the scenes to create a child meal plan, an adult meal plan, a combined meal plan, and a post-processing agent that formats the output.

The diagram below illustrates the full architecture of the event-driven multi-agent system.

Confluent Cloud's Data Stream Platform is used as a shared communication and orchestration channel for the agents. The user-facing part of the web application doesn't know anything about AI.

Project overview

The project is split into two applications. The web-application is a NextJS application that uses a standard three tier stack consisting of a frontend written in React, a backend in Node, and a MongoDB application database.

Kafka and Flink, running on Confluent Cloud, are used to move data around between services. The web application doesn't know anything about LLMs, Kafka, or Flink.

The agents application is a Python app that includes routes to the different agents and API endpoints called by Confluent to consume messages from Kafka topics. These API endpoints take care of all the AI magic to generate a meal plan and grocery list.

What you'll need

In order to set up and run the application, you need the following:

Node v22.5.1 or above
Python 3.10 or above
A Confluent Cloud account
A Claude API key
A LangChain API key

Getting set up

Get the starter code

In a terminal, clone the sample code to your project's working directory with the following command:

git clone https://github.com/thefalc/meal-planning-agent.git

Setting up MongoDB

In MongoDB create a database called meal_planner with the following collections:

meal_preferences - Stores your prefences for generating meals
weekly_meal_plans - Will contain the meal plan requests and results

Configure and run the Meal Planner AI web application

Go into your web-application folder and create a .env file with your MongoDB connection details.

MONGODB_URI='mongodb+srv://USER:PASSWORD@CLUSTER_URI/?retryWrites=true&w=majority&appName=REPLACE_ME'

Navigate into the web-application folder and run the application.

npm install
npm run dev

Go to http://localhost:3000 and try creating a meal plan request. If everything looks good, then continue with the setup.

Setting up Confluent Cloud

The meal planner uses Confluent Cloud to move and operate on data in real-time and handle the heavy lifting for communication between the agents.

Create the MongoDB meal planner request source connector

In order to kick start the agentic workflow, data from MongoDB needs to be published to Kafka. This can be done by creating a MongoDB source connector.

In Confluent Cloud, create a new connector.

Search for "mongodb" and select the MongoDB Atlas Source
Enter a topic prefix as meal-planner.input.request
In Kafka credentials, select Service account and use an existing or create a new one
In Authentication, enter your MongoDB connection details, the database name meal_planner and a collection name of weekly_meal_plans
Under Configuration, select JSON
For Sizing, leave the defaults and click Continue
Name the connector meal-planner-requests and click Continue

Create the HTTP sink connector to create the child meal plan

Now that data is flowing from the application database into the meal-planner.input.request.meal_planner.weekly_meal_plans topic, you now need to setup two HTTP sink connectors 8000 to create the child and adult meal plans.

Under Connectors, click + Add Connector
Search for "http" and select the HTTP Sink connector
Select the meal-planner.input.request.meal_planner.weekly_meal_plans topic
In Kafka credentials, select Service account and use you existing service account and click Continue
Enter the URL for where the child-preferences-agent endpoint is running under the agents folder. This will be similar to https://YOUR-PUBLIC-DOMAIN/api/child-preferences-agent. If running locally, you can use ngrok to create a publicly accessible URL. Click Continue
Under Configuration, select JSON and click Continue
For Sizing, leave the defaults and click Continue
Name the connector meal-planner-child-meal-plan and click Continue

Once the connector is created, under the Settings > Advanced configuration make sure the Request Body Format is set to json.

Repeat this same set of steps for the adult-preferences-agent endpoint.

Create the child meal plan topic

The child-preferences-agent endpoint publishes the child friendly meal plan to a Kafka topic called meal-planner.output.child-preferences.

In your Confluent Cloud account.

Go to your Kafka cluster and click on Topics in the sidebar.
Name the topic as meal-planner.output.child-preferences.
Set other configurations as needed, such as the number of partitions and replication factor, based on your requirements.
Go to Schema Registry
Click Add Schema and select meal-planner.output.child-preferences as the subject
Choose JSON Schema as the schema type
Paste the schema from below into the editor

{
  "properties": {
    "content": {
      "connect.index": 1,
      "oneOf": [
        {
          "type": "null"
        },
        {
          "type": "string"
        }
      ]
    },
    "request_id": {
      "connect.index": 0,
      "oneOf": [
        {
          "type": "null"
        },
        {
          "type": "string"
        }
      ]
    }
  },
  "title": "Record",
  "type": "object"
}

Save the schema

Create the adult meal plan topic

The adult-preferences-agent endpoint publishes the adult friendly meal plan to a Kafka topic called meal-planner.output.adult-preferences.

In your Confluent Cloud account.

Go to your Kafka cluster and click on Topics in the sidebar.
Name the topic as meal-planner.output.adult-preferences.
Set other configurations as needed, such as the number of partitions and replication factor, based on your requirements.
Go to Schema Registry
Click Add Schema and select meal-planner.output.adult-preferences as the subject
Choose JSON Schema as the schema type
Paste the schema from below into the editor

{
  "properties": {
    "content": {
      "connect.index": 1,
      "oneOf": [
        {
          "type": "null"
        },
        {
          "type": "string"
        }
      ]
    },
    "request_id": {
      "connect.index": 0,
      "oneOf": [
        {
          "type": "null"
        },
        {
          "type": "string"
        }
      ]
    }
  },
  "title": "Record",
  "type": "object"
}

Save the schema

Flink SQL setup to join the meal plans

Flink SQL is used to combine the child and adult meal plans into a single topic called meal-planner.output.joined-preferences. As events are written into this topic, the shared-preferences-agent is triggered to combine the meals plans into a single plan for the family.

To set this up, in your Confluent Cloud account.

In your Kafka cluster, go to the Stream processing tab
Click Create workspace
Create a table for storing both meal plans using the connection you created in the previous step

CREATE TABLE `default`.`dev-us-east-1-cluster`.`meal-planner.output.joined-preferences` (
    `request_id` STRING,
    `child_preference` STRING,
    `adult_preference` STRING
) WITH (
  'value.format' = 'json-registry'
);

Click Run

Both the meal-planner.output.child-preferences and meal-planner.output.adult-preferences topics are populating as meal plans are requested. We want to combine these independent processes together so we can start the shared agent meal planning service.

To do this, I'm populating the meal-planner.output.joined-preferences topic when meal-planner.output.child-preferences and meal-planner.output.adult-preferences can be joined based on the request_id.

To set this up, in the same workspace where you created the table, enter the following and click Run.

INSERT INTO `meal-planner.output.joined-preferences`
SELECT c.`request_id`, c.content as child_preference, a.content as adult_preference FROM `meal-planner.output.child-preferences` c
  JOIN `meal-planner.output.adult-preferences` a ON c.`request_id` = a.`request_id`;

The next step to configuring Confluent Cloud is to create another HTTP sink connector that will call the shared-preferences-agent API endpoint to create the shared meal plan.

Created the shared preference agent HTTP sink

In your Confluent Cloud account.

Under Connectors, click + Add Connector
Search for "http" and select the HTTP Sink connector
Select the meal-planner.output.joined-preferences topic
In Kafka credentials, select Service account and use you existing service account and click Continue
Enter the URL for where the shared-preferences-agent endpoint is running under the agents folder. This will be similar to https://YOUR-PUBLIC-DOMAIN/api/shared-preferences-agent. If running locally, you can use ngrok to create a publicly accessible URL. Click Continue
Under Configuration, select JSON_SR and click Continue
For Sizing, leave the defaults and click Continue
Name the connector meal-planner-shared-preferences and click Continue

Once the connector is created, under the Settings > Advanced configuration make sure the Request Body Format is set to json.

Create the shared meal plan topic

The shared-preferences-agent endpoint publishes the a combined meal plan to a Kafka topic called meal-planner.output.raw-complete-meal-plan.

In your Confluent Cloud account.

Go to your Kafka cluster and click on Topics in the sidebar.
Name the topic as meal-planner.output.raw-complete-meal-plan.
Set other configurations as needed, such as the number of partitions and replication factor, based on your requirements.
Go to Schema Registry
Click Add Schema and select meal-planner.output.raw-complete-meal-plan as the subject
Choose JSON Schema as the schema type
Paste the schema from below into the editor

{
  "properties": {
    "meal_plan": {
      "connect.index": 1,
      "oneOf": [
        {
          "type": "null"
        },
        {
          "type": "string"
        }
      ]
    },
    "request_id": {
      "connect.index": 0,
      "oneOf": [
        {
          "type": "null"
        },
        {
          "type": "string"
        }
      ]
    }
  },
  "title": "Record",
  "type": "object"
}

Save the schema

Created the format output agent HTTP sink

The format output agent takes the raw combined meal plan and formats it into JSON.

In your Confluent Cloud account.

Under Connectors, click + Add Connector
Search for "http" and select the HTTP Sink connector
Select the meal-planner.output.raw-complete-meal-plans topic
In Kafka credentials, select Service account and use you existing service account and click Continue
Enter the URL for where the format-output-agent endpoint is running under the agents folder. This will be similar to https://YOUR-PUBLIC-DOMAIN/api/format-output-agent. If running locally, you can use ngrok to create a publicly accessible URL. Click Continue
Under Configuration, select JSON_SR and click Continue
For Sizing, leave the defaults and click Continue
Name the connector meal-planner-format-output and click Continue

Once the connector is created, under the Settings > Advanced configuration make sure the Request Body Format is set to json.#### Created the shared preference agent HTTP sink.

Create the formatted output topic

The format-output-agent endpoint publishes the a combined meal plan and grocery list as JSON to a Kafka topic called meal-planner.output.formatted-complete-meal-plan.

In your Confluent Cloud account.

Go to your Kafka cluster and click on Topics in the sidebar.
Name the topic as meal-planner.output.formatted-complete-meal-plan.
Set other configurations as needed, such as the number of partitions and replication factor, based on your requirements.
Go to Schema Registry
Click Add Schema and select meal-planner.output.raw-complete-meal-plan as the subject
Choose JSON Schema as the schema type
Paste the schema from below into the editor

{
  "properties": {
    "meal_plan": {
      "connect.index": 1,
      "oneOf": [
        {
          "type": "null"
        },
        {
          "type": "string"
        }
      ]
    },
    "request_id": {
      "connect.index": 0,
      "oneOf": [
        {
          "type": "null"
        },
        {
          "type": "string"
        }
      ]
    }
  },
  "title": "Record",
  "type": "object"
}

Save the schema

Created the save output HTTP sink

The save meal plan endpoint takes the JSON output of the format output agent and writes it to MongoDB.

In your Confluent Cloud account.

Under Connectors, click + Add Connector
Search for "http" and select the HTTP Sink connector
Select the meal-planner.output.formatted-complete-meal-plan topic
In Kafka credentials, select Service account and use you existing service account and click Continue
Enter the URL for where the save-meal-plan endpoint is running under the agents folder. This will be similar to https://YOUR-PUBLIC-DOMAIN/api/save-meal-plan. If running locally, you can use ngrok to create a publicly accessible URL. Click Continue
Under Configuration, select JSON_SR and click Continue
For Sizing, leave the defaults and click Continue
Name the connector meal-planner-save-meal-plan and click Continue

Once the connector is created, under the Settings > Advanced configuration make sure the Request Body Format is set to json.#### Created the shared preference agent HTTP sink.

Run the application

In a terminal, navigate to your project directory. Run the app with the following command:

python -m venv env
source env/bin/activate
pip install -r requirements.txt
uvicorn app.main:app --reload

From your browser, navigate to http://localhost:3000 and you should see the meal planner AI app page.
Configure your settings.
Click Request a Meal Plan.
Wait for the agent flow to complete and click the entry to view the full details of the meal plan and grocery list.

Name		Name	Last commit message	Last commit date
Latest commit History 3 Commits
agents		agents
images		images
web-application		web-application
.gitignore		.gitignore
LICENSE		LICENSE
README.md		README.md

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

Multi-Agent AI Meal Planner

How it works

Project overview

What you'll need

Getting set up

Get the starter code

Setting up MongoDB

Configure and run the Meal Planner AI web application

Setting up Confluent Cloud

Create the MongoDB meal planner request source connector

Create the HTTP sink connector to create the child meal plan

Create the child meal plan topic

Create the adult meal plan topic

Flink SQL setup to join the meal plans

Created the shared preference agent HTTP sink

Create the shared meal plan topic

Created the format output agent HTTP sink

Create the formatted output topic

Created the save output HTTP sink

Run the application

About

Releases

Packages

Languages

License

thefalc/meal-planning-agent

Folders and files

Latest commit

History

Repository files navigation

Multi-Agent AI Meal Planner

How it works

Project overview

What you'll need

Getting set up

Get the starter code

Setting up MongoDB

Configure and run the Meal Planner AI web application

Setting up Confluent Cloud

Create the MongoDB meal planner request source connector

Create the HTTP sink connector to create the child meal plan

Create the child meal plan topic

Create the adult meal plan topic

Flink SQL setup to join the meal plans

Created the shared preference agent HTTP sink

Create the shared meal plan topic

Created the format output agent HTTP sink

Create the formatted output topic

Created the save output HTTP sink

Run the application

About

Resources

License

Stars

Watchers

Forks

Releases

Packages 0

Languages

Packages