🚀 AI Books MCP Server - Extend LLM context 15-60× via gravitational memory | Official MCP server for Claude Code & Anthropic

🧠 Extend LLM context windows up to 60× with perfect data integrity using gravitational memory compression for fast, universal compatibility.

The source codes for bachelor's thesis

Benchmarking RoPE scaling methods on TinyLlama-1.1B to measure perplexity degradation when extending context from 2048 to 8192 tokens. NTK-aware scaling preserves quality almost perfectly while linear scaling collapses.