About Akhbar
Project Overview
Akhbar (أخبار), meaning "news" in Arabic, is an experimental news aggregation and analysis project that combines modern web technologies with AI capabilities. We focus on providing innovative ways to interact with news content while respecting copyright and data protection.
Content Handling & Legal Compliance
Source Types
Content accessed through official news APIs with proper authentication and usage tracking.
Brief excerpts and summaries within fair use, always with attribution and links to original sources.
Direct links to articles with no content storage. Users can paste content for temporary AI analysis.
Only for sources that explicitly allow content reproduction (e.g., open licenses, press releases).
RSS Feed Integration
For sources providing RSS feeds, we track feed metadata and update frequencies while respecting each source's terms of service and rate limits. Content is processed according to the source type permissions above.
AI Interaction & Data Processing
AI Processing
AI interactions are processed using state-of-the-art language models, with flexibility to use different engines:
- Support for various AI engines (e.g., Ollama, LlamaCpp, LocalAI)
- Multiple model options (e.g., Gemma, Llama, Mistral)
- Local processing with no external API dependencies
- Content processed in-memory only
- No permanent storage of user-provided content
AI Personas
Customizable AI personas provide diverse interaction styles through advanced prompt engineering. Current examples include:
- Bored Teen: Casual, relatable explanations with modern references
- News Agent: Professional, journalistic analysis
- Sports Commentator: Energetic, engaging breakdowns
- Academic: Scholarly analysis with citations (coming soon)
- Data Analyst: Numbers-focused insights (coming soon)
All personas maintain appropriate respect for serious topics while providing their unique perspective. The system is designed to be easily extended with new personas.
Content Processing
Our content processing pipeline includes:
- Automatic sensitivity detection for appropriate response toning
- Topic classification and categorization
- Smart prompt templates that adapt to content type
- Response moderation and quality checks
System prompts and response guidelines are continuously refined to ensure engaging yet responsible AI interactions across all content types.
Important Disclaimers
Content & Copyright
- News content remains the intellectual property of respective organizations
- Links and summaries are provided for informational purposes only
- We do not store full article content without explicit permission
- User-provided content is processed temporarily and not stored
AI Interactions
- AI responses are generated locally using open-source models
- Responses may not be accurate or complete
- Users should verify information from reliable sources
- No guarantee of service availability or performance
Usage & Privacy
- This is an experimental project not intended for production use
- No user data is collected beyond basic operation logs
- API access logs are maintained for rate limiting only
- Feed monitoring respects source terms of service
Creator
Akhbar was created by Abdulkareem H. Ali as an experimental project exploring the intersection of news media and AI interaction while maintaining legal compliance and ethical content handling.