VSCode Extension for AI Code Assistance

This VSCode extension provides AI-assisted code suggestions based on user comments in supported programming languages.

Features

• You Choose LLM Model: You have flexibility to use any chat model from any provider — supports 5 providers as of today - OpenAI, AWS Bedrock, Google GenAI, Azure , Anthropic.
• Language Support: Any language you can use in VSCode.
• AI Suggestions: Fetches code suggestions based on user-defined questions in comments.
• Multi-line Comment Handling: Supports multi-line comments with customizable start and end tokens.
• Provider Overriding: Users can specify a different AI provider by including provider= in comments.
• Include File Content in Queries: Users can reference and include content from other files in their queries, useful for test cases creation ot code conversion or creating Readme -- See Ex. below.
• Explain and Review Code: Users can review the code in current active file or select a piece of code to get explanation.
• Bonus This provides a pay-as-you-go option, rather than monthly payments for various copilots.

Usage

1. Code Generation:

   • For single-line comments: Use the format // @! Your question (or any single-line comment format specific to your programming language).
   • For multi-line comments: Use the format (or ''' for Python):

     /* 
       Your question 
       @! */
     

   • Extension understands comment format for - js, ts, java, csharp, python and rust. For other languages it uses default of "//" for singleline and /* */ for multiline. These can be overriden in settings

2. Include External Files in AI Query (New Feature):

   • You can now reference and include content from other files in your AI query by using the include=<file> directive. This is particularly useful when you want the AI to consider code from multiple files, such as including the definition of functions or creating test cases or Readme based on file contents.

     // @! Write test cases for this function include=utils.js
     

3. Explain Code:

   • Select the code and run explain code command from context menu. Explanation is shown in side window. This uses default provider configured for extension.

4. Review Code:

   • Run review code command from context menu. Review is shown in side window. This uses default provider configured for extension.

5. Add Comment:

   • Run add comment command from context menu. Comment is added at top of selected code. This uses default provider configured for extension.

6. AI Interaction:

   • The extension captures the question, appends the code prior to the comment as context, and retrieves a code suggestion from the AI model.

7. Updating Suggestions:

   • After modifying a comment, press Enter at the last line to request a new code suggestion.

8. Customizing Provider:

   • Override the default AI provider by specifying them in your comment, like so:

     // @! Your question provider=yourProvider
     

Example Use Case: Creating Test Cases with the Include Feature or Convert to another language

Suppose you have a function defined in utils.js and you want to write test cases for it. You can use the include feature to pull the content from utils.js and get test case suggestions from the AI:

// @! Write unit tests for this function include=utils.js

Example Use Case: Code Conversion from JavaScript to Python

If you have JavaScript code in a file (e.g., example.js) and you want to convert it to Python, you can use the include feature to fetch the content from the file, in your python file (say example.py) and ask the AI for a conversion suggestion:

# @! Convert the JavaScript code to Python include=example.js

Configuration

The extension provides a variety of configuration options that can be set either through the settings menu or by editing the settings.json file in VSCode. These configurations are split into sections for common settings and provider-specific settings.

Common Settings

OpenAI Settings

Azure Settings

AWS Settings

Google Settings

Anthropic Settings

Comment Configuration for Other Languages