Hearbitz

Automatically ingests news articles from undisclosed sources and generates condensed summaries using an unspecified LLM, then applies one of three fixed editorial personas (Neutral, Progressive, Conservative) to reframe or filter the summary before audio conversion. The persona layer appears to operate as a post-summarization rewriting or filtering mechanism rather than prompt-level injection, though implementation details are not documented.

Converts summarized news text to natural-sounding audio using an unspecified TTS engine (likely Google Cloud TTS, Azure Speech, or ElevenLabs based on industry standards), then streams the audio through a web-based player with adjustable playback speed (range unknown, likely 0.75x–2.0x). The TTS engine and voice selection (gender, accent, language-specific voices) are not documented.

Allows users to select topics of interest (specific topics unknown — likely categories like Business, Technology, Politics, Sports, Health) during onboarding, then filters the news feed to show only articles matching those topics. The filtering mechanism is not documented — unclear if it uses keyword matching, semantic similarity, or editorial tagging. Feed refresh frequency and article selection algorithm are unknown.

Provides a browser-based audio player interface accessible at hearbitz.app with basic playback controls (play/pause, skip forward/backward, speed adjustment). The player is likely built using HTML5 audio element or a third-party player library (e.g., Plyr, JW Player), with no mention of advanced features like bookmarking, note-taking, or transcript search.

Ingests news articles from undisclosed sources (likely news APIs like NewsAPI, GNews, or partnerships with news outlets) and applies an unknown selection algorithm to choose which articles to include in daily briefings. The algorithm likely considers recency, topic relevance, and persona alignment, but specifics are not documented. No information on source diversity, editorial review, or fact-checking is provided.

Orchestrates a multi-stage pipeline: news ingestion → summarization → persona filtering → text-to-speech conversion → audio streaming. The pipeline is likely asynchronous (articles are pre-processed and cached rather than generated on-demand) to minimize latency, but specifics are unknown. No information on caching strategy, CDN usage, or real-time vs. batch processing is provided.

Offers free access to news summaries and audio briefings with an unspecified paywall or usage limit. The pricing page is not visible on the website, suggesting either a hidden freemium model (free tier with limited articles/minutes per day, paid tier for unlimited access) or a signup-wall that reveals pricing after account creation. No information on free tier limits, paid tier pricing, or upgrade triggers is provided.

Stores user preferences (topics, persona, playback speed) in an account system (likely a relational database with user authentication) to enable personalization across sessions and devices. The account system is required to access the service, but no information on data retention, privacy, or account recovery is provided. No mention of social login, single sign-on, or multi-device sync is made.

Provides a responsive web interface accessible at hearbitz.app that adapts to mobile, tablet, and desktop screen sizes. The interface is likely built using a modern web framework (React, Vue, or similar) with CSS media queries for responsive design. No information on accessibility (WCAG compliance), keyboard navigation, or screen reader support is provided.