It has now become possible to run local LLMs with a reasonable speed. There is a large selection of models available that can be used. There are many apps out there already that make it easy to run models locally:
It may not even be necessary to require local LLMs to be used. It would be possible to have support for external LLM APIs, should the user decide to go that route.
I suggest a plugin for CodeLite similar to GitHub Copilot's Visual Studio Code extension.
Possible Features:
Model / API selection and configuration
Code completion in code editor
AI Chat in IDE widget
Retrieval-augmented generation (RAG) using project files
Autonomous task completion using function calling
This feature is critical for CodeLite. It would significantly enhance the coding experience, reducing development time and increasing productivity. It would allow CodeLite to compete with other IDEs that either have this feature already or will have it in the near future.
Describe your idea in details
It has now become possible to run local LLMs with a reasonable speed. There is a large selection of models available that can be used. There are many apps out there already that make it easy to run models locally:
It may not even be necessary to require local LLMs to be used. It would be possible to have support for external LLM APIs, should the user decide to go that route.
I suggest a plugin for CodeLite similar to GitHub Copilot's Visual Studio Code extension.
Possible Features:
This feature is critical for CodeLite. It would significantly enhance the coding experience, reducing development time and increasing productivity. It would allow CodeLite to compete with other IDEs that either have this feature already or will have it in the near future.