GCF: A Token-Optimized Wire Format for Structured LLM Interactions

GCF is an AI-native wire format for structured data in LLM systems. 2,400+ evaluations across 11 models and 3 providers. 43 billion+ lossless round-trips across 5 formats. Two published papers.

Comprehension: 100% accuracy on every frontier model on standard workloads. 91.2% on structurally complex data where JSON drops to 53.4% and TOON to 68.2%. No model has ever been trained on GCF.

Token efficiency: 50-92% fewer tokens than JSON (50-69% per call, up to 92% with session deduplication). JSON overhead is 81% structural noise at scale; GCF overhead is 0.2%.

Tokenizer analysis: JSON's grammar symbols merge with common field names (id, name, type, value) in BPE tokenizer vocabularies on 50-63% of tested tokenizers. These are hardcoded, irrecoverable dictionary entries. GCF's pipe delimiter has 0% merge rate with field names. See companion paper: "Structural Ambiguity in JSON Tokenization" (DOI: 10.5281/zenodo.20789620).

Six implementations (Go, TypeScript, Python, Rust, Swift, Kotlin), 157 conformance fixtures, zero dependencies. Specification v3.1 Stable. Lossless across 17 serialization formats.