Articles & research
-
Does a Language Server Save Tokens for Coding Agents? A Measurement Methodology and Preliminary Study
Does semantic code retrieval actually reduce an agent’s token budget? This preliminary study compares LSP with lexical search using tokens-to-success at equal task success. The findings suggest that LSP does not consistently save tokens: its value depends on the task, model capability, and lexical noise.
The same study is discussed on the AgentConnect blog in Grep beats LSP? Why coding agents ignore your fancier tools.