Local LLM-driven code completion for Neovim. The idea is to leverage the ability of some LLMs to perform fill-in-the-middle (FIM) tasks, where you provide a prefix and a suffix as context for the completion.
Caution
This is a work in progress experiment and might not work as expected.
Boldo uses Ollama, make sure it's up and running with at least one of the supported models before continuing to the installation.
{
'lupeke/boldo.nvim',
config = function()
require('boldo').setup()
end
}Packer
use {
'lupeke/boldo.nvim',
config = function()
require('boldo').setup()
end
}:Boldo:BoldoModel <model>
<C-b>lwill trigger the completion. Use the defaultundo shortcut to discard the completion.
The following models are supported:
| model |
|---|
| qwen2.5-coder:14b |
| deepseek-coder:6.7b |
| qwen2.5-coder:7b |
| codellama:7b-code |
| codegemma:7b-code |
