# PARSER: Read in Parallel, Reason in Depth for Long-Context LLM Agents > Source: > Published: 2026-09-10 17:30:03+00:00 Sequential memory agents process long documents by reading chunks one after another while maintaining a compact memory state, coupling document traversal to reasoning depth. This coupling introduces sensitivity to evidence placement and ties inference latency linearly to document length. We introduc