Complete local HTTP endpoint parity

src/engine.rs

@@ -560,7 +560,7 @@ impl Generator {
         let max_context = self.executor.context() as usize;
         if tokens.len() >= max_context {
             return Err(format!(
-                "the conversation uses {} tokens; the configured context holds fewer than {max_context}",
+                "Prompt has {} tokens, but the configured context size is {max_context} tokens",
                 tokens.len()
             ));
         }