Content
# Purchasing Concierge A2A Demo
This demo shows how to enable A2A (Agent2Agent) protocol communication between purchasing concierge agent with the remote pizza and burger seller agents using A2A Python SDK. Burger and Pizza seller agents is a independent agent that can be run on different server with different frameworks, in this demo example we, burger agent is built on top of Crew AI and pizza agent is built on top of LangGraph.
Detailed tutorial about this repo: [Getting Started with Agent2Agent (A2A) Protocol: A Purchasing Concierge and Remote Seller Agent Interactions with Gemini on Cloud Run](https://codelabs.developers.google.com/intro-a2a-purchasing-concierge?utm_campaign=CDR_0x6a71b73a_default_b415667894&utm_medium=external&utm_source=blog)
## Prerequisites
- If you are executing this project from your local IDE, Login to Gcloud using CLI with the following command :
```shell
gcloud auth application-default login
```
- Enable the following APIs
```shell
gcloud services enable aiplatform.googleapis.com
```
- Install [uv](https://docs.astral.sh/uv/getting-started/installation/) dependencies and prepare the python env
```shell
curl -LsSf https://astral.sh/uv/install.sh | sh
uv python install 3.12
uv sync --frozen
```
## How to Run
First, we need to run the remote seller agents. We have two remote seller agents, one is burger agent and the other is pizza agent. We need to run them separately. These agents will serve the A2A Server
### Run the Burger Agent - Locally
1. Copy the `remote_seller_agents/burger_agent/.env.example` to `remote_seller_agents/burger_agent/.env`.
2. Fill in the required environment variables in the `.env` file. Substitute `GOOGLE_CLOUD_PROJECT` with your Google Cloud Project ID.
```bash
GOOGLE_CLOUD_LOCATION=us-central1
GOOGLE_CLOUD_PROJECT={your-project-id}
```
3. Run the burger agent.
```bash
cd remote_seller_agents/burger_agent
uv sync --frozen
uv run .
```
4. It will run on `http://localhost:10001`
### Run the Pizza Agent - Locally
1. Copy the `remote_seller_agents/pizza_agent/.env.example` to `remote_seller_agents/pizza_agent/.env`.
2. Fill in the required environment variables in the `.env` file. Substitute `GOOGLE_CLOUD_PROJECT` with your Google Cloud Project ID.
```bash
GOOGLE_CLOUD_LOCATION=us-central1
GOOGLE_CLOUD_PROJECT={your-project-id}
```
3. Run the pizza agent.
```bash
cd remote_seller_agents/pizza_agent
uv sync --frozen
uv run .
```
4. It will run on `http://localhost:10000`
### Run the Purchasing Concierge Agent - Locally
Finally, we can run our A2A client capabilities owned by purchasing concierge agent.
1. Go back to demo root directory ( where `purchasing_concierge` directory is located )
2. Copy the `purchasing_concierge/.env.example` to `purchasing_concierge/.env`.
3. Fill in the required environment variables in the `.env` file. Substitute `GOOGLE_CLOUD_PROJECT` with your Google Cloud Project ID.
And fill in the `PIZZA_SELLER_AGENT_URL` and `BURGER_SELLER_AGENT_URL` with the URL of the remote seller agents.
```bash
PIZZA_SELLER_AGENT_URL=http://localhost:10000
BURGER_SELLER_AGENT_URL=http://localhost:10001
GOOGLE_GENAI_USE_VERTEXAI=TRUE
GOOGLE_CLOUD_PROJECT={your-project-id}
GOOGLE_CLOUD_LOCATION=us-central1
```
4. Run the purchasing concierge agent with the adk web dev UI
```bash
uv sync --frozen
uv run adk web
```
## Deployment
### Deploy the Burger Agent - Cloud Run
Run the following command
```bash
gcloud run deploy burger-agent \
--source remote_seller_agents/burger_agent \
--port=8080 \
--allow-unauthenticated \
--min 1 \
--region us-central1 \
--update-env-vars GOOGLE_CLOUD_LOCATION=us-central1 \
--update-env-vars GOOGLE_CLOUD_PROJECT={your-project-id}
gcloud run deploy burger-agent-east \
--source remote_seller_agents/burger_agent \
--port=8080 \
--allow-unauthenticated \
--min 1 \
--region us-east4 \
--update-env-vars GOOGLE_CLOUD_LOCATION=us-east4 \
--update-env-vars GOOGLE_CLOUD_PROJECT=mahen-projects
```
### Deploy the Pizza Agent - Cloud Run
Run the following command
```bash
gcloud run deploy pizza-agent-east4 \
--source remote_seller_agents/pizza_agent \
--port=8080 \
--allow-unauthenticated \
--min 1 \
--region us-east4 \
--update-env-vars GOOGLE_CLOUD_LOCATION=us-east4 \
--update-env-vars GOOGLE_CLOUD_PROJECT=mahen-projects
```
### Deploy Purchasing Concierge Agent - Agent Engine
1. Create the staging bucket first
```bash
gcloud storage buckets create gs://purchasing-concierge-{your-project-id} --location=us-central1
```
2. Copy the `.env.example` to `.env`.
3. Fill in the required environment variables in the `.env` file. Substitute `GOOGLE_CLOUD_PROJECT` with your Google Cloud Project ID.
```bash
GOOGLE_GENAI_USE_VERTEXAI=TRUE
GOOGLE_CLOUD_PROJECT={your-project-id}
GOOGLE_CLOUD_LOCATION=us-central1
STAGING_BUCKET=gs://purchasing-concierge-{your-project-id}
PIZZA_SELLER_AGENT_URL={your-pizza-agent-url}
BURGER_SELLER_AGENT_URL={your-burger-agent-url}
```
4. Deploy the purchasing concierge agent to agent engine
```bash
uv sync --frozen
uv run deploy_to_agent_engine.py
```
### Run the Chat Interface to Connect to Agent Engine
1. Update the `.env` file with the `AGENT_ENGINE_RESOURCE_NAME` which obtained from the previous step.
2. Run the Gradio app
```bash
uv sync --frozen
uv run purchasing_concierge_ui.py
```
