Chat
Initialize any web chat with your code, hands-free!
Chat conversations are only a construct of product interfaces, they hurt the quality of responses from the model and once your context is "poisoned" it will not recover. Whenever you're not satisfied with a reponse, the best practice is to alawys refine your initial instructions and re-initialize a chat.
Generated prompt structure
- General
- Code Completions
[INSTRUCTIONS]
<files>
<text title="">...</text>
// other websites...
<file path="...">...</file>
// other files...
</files>
[INSTRUCTIONS]
For model's greater adherence to instructions, they're palced at both ends of the prompt.
🔗 OpenAI Cookbook/Prompt Organization
<files>
<text title="...">...</text>
// other websites...
<file path="...">...</file>
// other files excl. current editor...
<file path="..."> // current editor
...
<missing text> // cursor position
...
</file>
</files>
Find correct replacement text for the <missing text> symbol. Correctly formatted response begins with a code block containing replacement text end then proceeds with explanation. Always refer to symbol "<missing_text>" as "cursor position" and "replacement" as "completion".
Presets
Create general or purpose-tailored chat configurations.
Available commands
Code Web Chat: Web Chat
Opens a chat session in your browser with the current context and prompt.
Code Web Chat: Web Chat with...
Lets you select which AI platform to use for your chat session.
Code Web Chat: Chat to Clipboard
Instead of opening a chat directly, copies the context and prompt to your clipboard for manual pasting.
Settings
Code Web Chat: Edit Format Instructions Truncated
Style instructions for chat responses when using truncated format.
Code Web Chat: Edit Format Instructions Whole
Style instructions for chat responses when showing complete files.
Code Web Chat: Edit Format Instructions Diff
Whenever proposing a file use the markdown code block syntax. Each code block should be a diff patch.