document stance & supporting info on AI / LLMs

[ixxie]

In Github issues and discussions, the Helix team has made their position on AI more or less clear; however, the information is scattered, and since the question keeps coming back, a dedicated section of the docs would be helpful.

The section would:

• Outline the team's position on LLM integration
• Track relevant issues & PRs working towards relevant features
• Suggest complementary tooling and link to it

If the maintainers agree with this idea, I would be happy to take a stab at a PR.

Position on LLM integration

From various issues and discussions, as far as I can tell, the position is:

1. LLM integrations will not be made part of Helix core, but rather delegated to extensions
2. Currently, you can use LLM LSPs to facilitate in-file code suggestions or use complementary CLI tooling
3. An upcoming plugin system would allow third-party developers to build more sophisticated LLM extensions for Helix

Issues & PRs for relevant features

• Allow configuring LSP for any language
• larger suggestions (textDocument/inlineCompletion)
• Steel based plugin system

Complementary tooling

LLM LSPs

• helix-chatgpt
• lsp-ai
• llm-lsp

CLI tools

Some CLI tools for working with LLMs can be interesting, since they can be used in multiplexing in tandem with Helix:

• aider
• aichat

[pascalkuthe]

The wiki is free to edit for all so you can add a page there. Other third party integrations are alsonfocumented there. LLM and AI are not part of the core editor so they don't belong in the docs.

[ixxie]

Well, this makes sense I guess.

However, I never use the wiki really, since GitHub Wikis are clunky and hard to navigate. So I'm not super inclined to contribute to it, since I feel like it probably wouldn't help anybody much.

[shortc]

@the-mikedavis If you don't mind, what is the reason for the downvote on this?

[sdcuyahoga]

the helix devs are notoriously insufferable