What It Does

This recipe builds a “newsroom” of cooperating agents that produces a personal daily briefing. Following the CrewAI pattern, each agent has a role, goal, and backstory: a researcher gathers stories on the topics you care about, an editor filters and ranks them, and a writer composes a concise briefing in your preferred format. CrewAI orchestrates the hand-offs between them so the output is a finished digest rather than raw search results.

The Stack

• CrewAI — the multi-agent framework that orchestrates role-playing agents and the sequential tasks they pass between each other.
• A search/scrape tool — CrewAI’s built-in tools (e.g. a web search or scraping tool) feed the researcher agent live stories.
• An LLM backend — CrewAI is model-agnostic; point it at a free or low-cost provider such as GitHub Models or DeepSeek Platform rather than a paid key.

Step-by-Step Outline

1. Scaffold the crew — crewai create (or follow “Build your first crew”) to get the agents/tasks YAML and a runnable project.
2. Define agents — a Researcher (gather), an Editor (filter/rank), and a Writer (compose), each with role/goal/backstory.
3. Define tasks — research today’s topics → curate the shortlist → write the briefing, chained sequentially.
4. Wire tools + model — attach a search/scrape tool to the researcher and set the crew’s LLM to your chosen provider.
5. Run + deliver — kick off the crew; pipe the final briefing to email, a file, or a chat channel.

Source

Automating a Newsroom Crew with AI Agents and CrewAI (built on the official crewAI-examples and Build your First CrewAI Agents)