@@ -24,9 +24,9 @@ Works with VS Code family of editors (Cursor, Antigravity, VSCodium, etc.).
 
 **LLMs are brilliant, yet mindless pattern matchers.** Because they lack properties similar to mental models, they can't _imagine_ how the correct output should look like _before_ their context is filled with files.
 
-Coding agents address this limitation via tool calling—although keyword search may work, it requires instructions complex enough to fully understand your intent.
+Coding agents address this limitation via tool calling—although keyword search may work, it requires instructions complex enough to fully understand your intent. For complex codebases, the process may run for hours and still yield uncertain results.
 
-**Code Web Chat proposes an alternative approach** where the model is guided with files included directly in the prompt. This allows simpler instructions and unlocks accuracy that can't get any better!
+**CWC proposes an alternative, more mindful 🧠 approach through context engineering.** Pick files and be certain what the model gets to see, type simpler instructions, and get edits immediately with accuracy that can't get any better!
 
 With files sorted by update recency, CWC helps utilize **[prompt caching](https://platform.openai.com/docs/guides/prompt-caching) across tasks**—consistently **save up to 90% on input tokens**, while reserving the more expensive output tokens for task-focused reasoning and code changes you're looking for.