Ollama support

[peperunas]

LLMs are more and more commonly used to provide coding support. Ollama is a great project that acts as a middleware to use local LLMs.

Adding support to it would be great.

Website: https://ollama.ai

[David-Else]

You can use https://github.com/sigoden/aichat and integrate it to some degree, especially if you are using a terminal multiplexer or a terminal with split windows.

Support LLMs

• OpenAI: gpt-3.5/gpt-4/gpt-4-vision
• Gemini: gemini-pro/gemini-pro-vision/gemini-ultra
• LocalAI: opensource LLMs and other openai-compatible LLMs
• Ollama: opensource LLMs
• Azure-OpenAI: user deployed gpt-3.5/gpt-4
• Ernie: ernie-bot-turbo/ernie-bot/ernie-bot-8k/ernie-bot-4
• Qianwen: qwen-turbo/qwen-plus/qwen-max/qwen-max-longcontext/qwen-vl-plus

[peperunas]

I meant an "LSP"-like implementation. A-la https://continue.dev

[tirithen]

This is the main feature I'm missing in helix, adding gpt support via an LSP and locally run models with ollama sounds like a great way to preserve the privacy while enabling helix to use gpt models.

[ikks]

Take a look at this one. leona/helix-gpt#42 . Seems that ollama is available for helix-gpt