Hosted agent with Amazon Bedrock AgentCore

This example wraps an agent in the entrypoint that Amazon Bedrock AgentCore Runtime expects. The service supplies the runtime around the agent, including session isolation, identity, and observability, so the code only declares how a request becomes a response.

Requirements

• Python 3.10+
• AWS credentials configured for a region where Amazon Bedrock is available
• Model access granted for the model identifier used in the script

Steps for running this example in the shell

1. Install dependencies:

python -m venv .venv

# macOS/Linux:
source .venv/bin/activate

# Windows Command Prompt:
.venv\Scripts\activate.bat

# Windows PowerShell:
.venv\Scripts\Activate.ps1

pip install -r requirements.txt

2. Run the agent locally:

python hosted_agent.py

3. Send a request to the local endpoint:

curl -X POST http://localhost:8080/invocations -H "Content-Type: application/json" -d "{\"prompt\": \"Explain context engineering in one sentence.\"}"

Output

The local server answers on /invocations with the model reply wrapped in a result field. The same entrypoint is what AgentCore Runtime invokes once the agent is deployed, so no code changes are needed between local runs and the managed runtime.