Search code, repositories, users, issues, pull requests...
We found that often with AI code assistants, they just don't seem smart enough to generate the correct code in the context required. Examples include
The thing is, all of these are not outside the capabilities of modern LLMs - with the right information in the context window, all these issues can be resolved.
With our initial release, we've solved multi-file support, and with subsequent releases we hope to solve the other problems as well.
Right now, we have a base of features that we believe make a good foundation for the developer experience that we expect from an AI code assistant.
We have several ideas of how we can improve the developer experience, and we need your feedback and votes in order to better shape the development of this product.
