Daily Sky Haiku: Building an Autonomous AI Poetry Agent with Amazon Bedrock

Описание и применение

Автономный генератор хайку, который ежедневно получает погоду, генерирует стихотворение с помощью Amazon Nova Lite и публикует его на сайте без пользовательского ввода.

Задача

Create a fully autonomous daily haiku generator that pulls real‑world weather data and publishes a new poem each morning without user intervention.

Результат

The system autonomously produces a weather‑inspired haiku each morning, stores it in S3, and serves it via a public website, eliminating manual prompts.

Рабочий процесс

• Schedule EventBridge rule to trigger daily at 7:00 AM IST (cron(30 1 * * ? *)).
• AWS Lambda function retrieves current weather for Kochi from Open‑Meteo API.
• Lambda constructs a prompt incorporating weather data and sends it to Amazon Bedrock (Nova Lite) to generate a three‑line haiku.
• Lambda combines the poem with location, temperature, humidity, weather conditions, and timestamp into a JSON object.
• Lambda writes the JSON to Amazon S3 as latest.json.
• Netlify‑hosted frontend fetches latest.json from S3 and displays the haiku on the public website.

Доказательства

Built using Amazon Bedrock, AWS Lambda, Amazon EventBridge, and Amazon S3, the application automatically retrieves weather information, generates a poem using Amazon Nova Lite, and publishes the latest result to a public website.

Источник

EventBridge Triggers the Workflow
I created an EventBridge scheduled rule named daily-sky-haiku-morning.
The schedule uses this cron expression: cron(30 1 * * ? *)

Источник

Lambda retrieves the current weather information for Kochi using the Open‑Meteo API.

Источник

Lambda prepares a creative prompt using the retrieved weather information and sends it to Amazon Bedrock.
I use Amazon Nova Lite to generate a weather-themed poem consisting of three lines.

Источник

The generated result is stored in Amazon S3 as a JSON object named: latest.json

Источник

The public frontend, deployed on Netlify, reads the stored JSON file and displays the latest haiku.

Источник