Agently Daily News Collector

Report Abuse

Basic Information

Agently Daily News Collector is an open-source, LLM-powered workflow showcase that automates the collection, selection, summarization and output of topical news. It is built on the Agently AI application development framework and is implemented as a Python project. A user provides a topic via a prompt and the agent pipeline generates an outline, searches for relevant items, picks and summarizes articles, compiles columns and produces a final markdown file (and example PDF exports). Configuration is driven by a SETTINGS.yaml file where API keys and model settings are placed. The repository includes example output files and shows runtime logs that record outline generation, search counts, picked items and summarization steps. Installation of dependencies and running the main Python script are documented in the README.

Links

Categorization

App Details

Features
The project orchestrates agent tasks to create a curated news digest. It uses the Agently framework for agent-based workflows and relies on duckduckgo-search for web search, BeautifulSoup for HTML parsing and PyYAML for configuration. The pipeline generates an outline, performs search queries, ranks and picks candidate items, produces summaries and recommendation comments, and formats results into markdown and optional PDF outputs. Runtime logging shows steps such as outline generation, search hit counts, chosen articles and summarization status. Users configure model keys and options in SETTINGS.yaml. The README documents install steps with pip and how to start the generator using a Python entry script. Example generated files are included to illustrate output structure.
Use Cases
This repository provides a ready-made automation for producing daily or topic-focused news digests without manual curation. By accepting a simple topic prompt it saves time in researching, filtering and summarizing news, producing a structured markdown report suitable for distribution or archiving. It is configurable via a SETTINGS.yaml file so teams can swap models, add API keys and adjust search behavior. The logged output gives transparency into pipeline stages and helps troubleshoot or refine prompts. Example markdown and PDF outputs demonstrate how curated content is organized, making it useful for researchers, communicators or anyone who needs repeatable, automated news collection workflows.

Please fill the required fields*