Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Provider configuration migration guide

Jump to bottom
yetone edited this page Jun 3, 2025 · 4 revisions

Due to the previous unscientific design of provider configuration in avante.nvim, some issues arose: such as not knowing how to list all providers and being unable to distinguish between request body and regular provider configuration. Therefore, the configuration items for providers have been restructured:

  1. The configurations for all built-in providers have been moved from the top level of all configurations to under the providers field.
  2. The configurations for all custom providers have been moved from vendors to under the providers field.
  3. All request body fields (such as temperature, max_tokens, max_completion_tokens, reasoning_effort, options for ollama provider) of a provider have been moved from the top level of their respective provider's configuration to the extra_request_body field within that provider's configuration.

Migration Example:

For instance, if your previous configuration was as follows (using lazy.nvim as an example):

{
  "yetone/avante.nvim",
  --- other configuration items ...
  opts = {
    --- other configuration items ...
    openai = {
      endpoint = "https://api.openai.com/v1",
      model = "gpt-4o", -- your desired model (or use gpt-4o, etc.)
      timeout = 30000, -- Timeout in milliseconds, increase this for reasoning models
      temperature = 0,
      max_completion_tokens = 8192, -- Increase this to include reasoning tokens (for reasoning models)
      reasoning_effort = "medium", -- low|medium|high, only used for reasoning models
    },
    ollama = {
      endpoint = "http://127.0.0.1:11434",
      timeout = 30000, -- Timeout in milliseconds
      options = {
        temperature = 0.75,
        num_ctx = 20480,
        keep_alive = "5m",
      },
    },
    vendors = {
      groq = {
        __inherited_from = 'openai',
        api_key_name = 'GROQ_API_KEY',
        endpoint = 'https://api.groq.com/openai/v1/',
        model = 'llama-3.3-70b-versatile',
        temperature = 1,
        max_tokens = 32768, -- remember to increase this value, otherwise it will stop generating halfway
        disable_tools = true,
      },
    },
  },
  --- other configuration items
}

You need to migrate it like this:

{
  "yetone/avante.nvim",
  --- other configuration items ...
  opts = {
    --- other configuration items ...
    providers = {
      openai = {
        endpoint = "https://api.openai.com/v1",
        model = "gpt-4o", -- your desired model (or use gpt-4o, etc.)
        timeout = 30000, -- Timeout in milliseconds, increase this for reasoning models
        extra_request_body = {
          temperature = 0,
          max_completion_tokens = 8192, -- Increase this to include reasoning tokens (for reasoning models)
          reasoning_effort = "medium", -- low|medium|high, only used for reasoning models
        },
      },
      ollama = {
        endpoint = "http://127.0.0.1:11434",
        timeout = 30000, -- Timeout in milliseconds
        extra_request_body = {
          options = {
            temperature = 0.75,
            num_ctx = 20480,
            keep_alive = "5m",
          },
        },
      },
      groq = {
        __inherited_from = 'openai',
        api_key_name = 'GROQ_API_KEY',
        endpoint = 'https://api.groq.com/openai/v1/',
        model = 'llama-3.3-70b-versatile',
        disable_tools = true,
        extra_request_body = {
          temperature = 1,
          max_tokens = 32768, -- remember to increase this value, otherwise it will stop generating halfway
        },
      },
    },
  },
  --- other configuration items
}
Clone this wiki locally