About Blerbz

Rankings combine engagement rate (15%), velocity (15%), viral lift (15%), freshness (12%), source reputation (7%), and additional factors. Diversity caps prevent one publisher from crowding the list. Scores update on a frequent cadence; APIs can return breakdowns where exposed.

The catalog spans major wires and national papers, technology and business outlets, regional news, and selected community sources. Reputation tiers (for example up to ~1.0 for top-tier wires and nationals) feed the ranking; community sources sit lower but still surface when velocity demands it.

• Top tier — e.g. NYT, BBC, Reuters, AP, major nationals.
• Broad & tech — Large broadcasters, tech and business publications.
• Regional & specialty — Local and industry verticals.
• Community — e.g. Reddit, Hacker News (with lower base reputation).

• Web — Next.js 16 (App Router), React 19, CSS modules and shared theme tokens.
• API — Next.js routes for product APIs; NUCLEUZ (Fastify, TypeScript) for aggregation, caching, and v2 REST.
• Data & cache — PostgreSQL, Redis / Upstash / KV-style stores, in-memory hot cache, request coalescing.
• AI & NLP — Grok-family models for enrichment, plus NLP for topics and entities; BLERBZ Labs adds video pipelines (e.g. Veo, FILMZ/LTX where configured).
• Analytics warehouse — dbt models for internal analytics.

v2 endpoints are prefixed with /api/v2/: documents (including aliases and ETags), feeds (latest, trending, blerbz100, topic, source), and sync for incremental consumption. Authentication uses API keys; responses include standard rate-limit headers. Run NUCLEUZ integrated with the app or as a standalone service depending on deployment.

Focuz and in-app assistants use AI to summarize and answer with news context. STREAM turns stories into short video. BLERBZ Labs (Mission Control) hosts production tools—clips, reframes, soundtracks, and more—behind the /labs experience and related proxies. Outputs are machine-generated: always verify against primary sources.