x-retrieval-systems

What problem does it solve?

This skill analyzes X's multi-stage retrieval architecture, explaining how the system narrows hundreds of millions of tweets to a small candidate pool and determines what is eligible to be ranked.

Core Features & Use Cases

• Decodes In-Network sourcing and Out-of-Network discovery to show how content is surfaced.
• Explains Earlybird indexing (Realtime, Protected, Archive) and Phoenix-based embedding retrieval for candidate generation.
• Analyzes latency and concurrency using a single-writer/multi-reader model to achieve sub-second results at scale.

Quick Start

Ask the AI to outline how Earlybird shards the in-network index and how Phoenix uses embedding-based discovery to surface candidate tweets.