FeedCraft is a lightweight, open-source RSS middleware designed to simplify the processing of RSS feeds. It caters to users and developers who need to enrich, filter, or transform RSS content, offering a powerful yet easy-to-use solution for managing information streams. The primary benefit is the ability to automate complex feed manipulations, including AI-driven summarization, translation, and content extraction, directly within an RSS workflow.

How It Works

FeedCraft operates by processing RSS feeds through a series of configurable steps called "AtomCrafts." These crafts can perform actions like extracting full text from articles, translating titles or content using AI, generating summaries, or filtering out unwanted content like marketing articles. Users can chain multiple AtomCrafts into "FlowCrafts" for complex pipelines or define "Recipes" to apply specific processing sequences to particular RSS sources, creating custom, enriched RSS feeds.

Quick Start & Requirements

• Primary install / run command: Deployment is recommended via Docker. A minimal docker-compose example is provided in the README.
• Non-default prerequisites and dependencies: Requires a running Redis instance, a browserless/chrome instance (for advanced full-text extraction), and access to an OpenAI-compatible LLM API (requiring API base URL, API key, model name, and type).
• Demo: A demo site is available at https://feed-craft.colinx.one for quick experience.

Highlighted Details

• AI Integration: Leverages OpenAI-compatible LLMs for advanced features like article summarization