Perplexity AI Productionizes Query-Aware Compression to Cut Context Tokens 70%

We've productionized query-aware compression for faster, cleaner, more-accurate search. Better context is better than more context. Our system cuts context tokens up to 70% while improving answer quality.