Skip to content

Commit ebe5962

Browse files
committed
Update README to improve clarity and emphasize prompt caching benefits
1 parent 23ec7a5 commit ebe5962

1 file changed

Lines changed: 5 additions & 3 deletions

File tree

‎README.md‎

Lines changed: 5 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -21,9 +21,11 @@ Autofill for chatbots. Works with VS Code, Cursor, Antigravity, and others.
2121

2222
**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 relevant files.
2323

24-
To address this limitation, **Code Web Chat lets you guide the model with context**. Files are intelligently sorted in a way prioritizing [prompt caching](https://platform.openai.com/docs/guides/prompt-caching) across tasks—put less strain on the planet and save up to 90% on input tokens.
24+
To address this limitation, **Code Web Chat lets you guide the model with context**. With that come big wins—instructions can be simple, and accuracy can't get any better!
2525

26-
**Messages are simple**, constructed from selected files, chosen edit format, and prompt:
26+
For efficiency, files are sorted in a way prioritizing **[prompt caching](https://platform.openai.com/docs/guides/prompt-caching) across tasks**—put less strain on the planet 🌱, and save up to 90% on input tokens!
27+
28+
**Example message:**
2729

2830
```
2931
<system>
@@ -41,7 +43,7 @@ Whenever showing a new, updated, renamed or deleted file, provide a brief explan
4143
Implement a subtract function.
4244
```
4345

44-
Once the response is ready, all changes can be integrated in a fully revertible way.
46+
Once the response is ready, changes can be integrated in a fully revertible way.
4547

4648
## Browser integration
4749

0 commit comments

Comments
 (0)