Does TabbyML plan on supporting an in-editor Copilot Chat?

[bendavis78]

I use CopilotChat.nvim in addition to Copilot.nvim for completion. I find there are many times where completion or comment prompts won't do.

For example

• "Please explain this function/code block/etc"
• "What are some different approaches to implement this? What are the pros/cons"
• (after selecting a line) "I'm seeing the following error on this line"
• Ability to quickly select & insert changes from the chat
• Ability to provide context, eg @filename
• Ability to fix a diagnostic problem from the editor w/ a single keystroke (eg, typing warning, etc)
• Ability to add customized prompts for certain specialized tasks to quickly launch explain/review/fixes

[wsxiaoys]

it's already supported in vscode

You'll need to set --chat-model to enable the feature, see https://tabby.tabbyml.com/docs/quick-start/installation/docker/ for an example.

[bendavis78]

Sorry I hit enter too early on my post. Thanks for the quick reply though! Are these features supported in nvim?

• "Please explain this function/code block/etc"
• "What are some different approaches to implement this? What are the pros/cons"
• (after selecting a line) "I'm seeing the following error on this line"
• Ability to quickly select & insert changes from the chat
• Ability to provide context, eg @filename
• Ability to fix a diagnostic problem from the editor w/ a single keystroke (eg, typing warning, etc)
• Ability to add customized prompts for certain specialized tasks to quickly launch explain/review/fixes

[wsxiaoys]

Instead of reimplementing it in Tabby, the best approach is likely to have CopilotChat.nvim support a backend other than GitHub Copilot itself. (Which shouldn't be too challenging, as it appears that only the chat/completions and /embeddings endpoints are used).