UglyFeed is an application for retrieving, aggregating, filtering, and rewriting RSS feeds using Large Language Models (LLMs). It targets users interested in content curation, automated content generation, and personalized news experiences, offering enriched RSS feeds with AI-powered summaries and translations.

How It Works

UglyFeed leverages LLMs to process RSS feed content. It retrieves feeds, aggregates items by similarity, and then uses LLM APIs (OpenAI, Ollama, Groq, Anthropic) to rewrite, summarize, or translate the content based on user-defined prompts. The rewritten content is saved to JSON and can be converted into a valid RSS feed, served via an HTTP server, or deployed to GitHub/GitLab.

Quick Start & Requirements

• Install: Docker is the primary installation method.
• Prerequisites: Internet connection, Docker, an LLM API key (OpenAI, Ollama, Groq, Anthropic), and an RSS reader.
• Configuration: Requires editing config.yaml or using the Streamlit web UI to specify feeds and LLM details.
• Docs: Extended documentation is available.

Highlighted Details

• Supports multiple LLM providers and models, including local Ollama deployments.
• Offers features like content filtering, post-moderation, and translation.
• Can serve generated feeds via HTTP or deploy them to GitHub/GitLab for CDN access.
• Includes a Streamlit-based web UI for easier management and content evaluation.

Maintenance & Community

The project is maintained by fabriziosalmi. Further community engagement details are not explicitly provided in the README.

Licensing & Compatibility

• License: AGPL3 License.
• Compatibility: The AGPL3 license is a strong copyleft license, requiring derivative works to also be open-sourced under AGPL3. This may impose restrictions on linking with closed-source applications.

Limitations & Caveats

The project acknowledges the potential for misuse in creating misleading content and advises responsible, ethical use. The AGPL3 license may restrict integration into proprietary software.