feat(readme): add GitHub Copilot provider docs and config defaults

- document COPILOT_TOKEN env var and add Copilot instructions and example
- mention copilot in provider list and README features
- add copilot provider default config and reasoning_effort docs in config.lua
- send reasoning_effort in copilot provider API payload and clarify openai handling

Author: aweis89

README.md

@@ -15,7 +15,7 @@ perfect commit message.
 
 ## Features
 
-- 🤖 Automatically generates commit messages using Gemini, OpenAI, or Anthropic APIs
+- 🤖 Automatically generates commit messages using Gemini, OpenAI, Anthropic, or GitHub Copilot APIs
   when you run `git commit -v`
 - 🎯 Works from terminal or within Neovim (using vim-fugitive)
 - 🤝 Non-intrusive - if you start typing, AI suggestions are added as comments instead
@@ -93,6 +93,12 @@ export OPENAI_API_KEY="your-api-key-here"
 
 **For Anthropic:**
 
+```bash
+export COPILOT_TOKEN="your-github-copilot-token-here"
+```
+
+**For Anthropic:**
+
 ```bash
 export ANTHROPIC_API_KEY="your-api-key-here"
 ```
@@ -110,7 +116,7 @@ require("ai_commit_msg").setup({
   -- Enable/disable the plugin
   enabled = true,
 
-  -- AI provider to use ("gemini", "openai", or "anthropic")
+  -- AI provider to use ("gemini", "openai", "anthropic", or "copilot")
   provider = "gemini",
 
   -- Whether to prompt for push after commit
@@ -194,6 +200,21 @@ require("ai_commit_msg").setup({
 
 ### Switch to Anthropic Claude
 
+### Switch to GitHub Copilot
+
+```lua
+require("ai_commit_msg").setup({
+  provider = "copilot",
+  providers = {
+    copilot = {
+      model = "gpt-5-mini",
+    },
+  },
+})
+```
+
+### Switch to Anthropic Claude
+
 ```lua
 require("ai_commit_msg").setup({
   provider = "anthropic",
@@ -320,6 +341,7 @@ git config --global core.editor nvim
   - Gemini: Set `GEMINI_API_KEY` environment variable (default, best value)
   - OpenAI: Set `OPENAI_API_KEY` environment variable
   - Anthropic: Set `ANTHROPIC_API_KEY` environment variable
+  - GitHub Copilot: Set `COPILOT_TOKEN` environment variable
 - Git
 - curl (for making API requests)
 

lua/ai_commit_msg/config.lua

@@ -10,7 +10,11 @@ local DEFAULT_SYSTEM_PROMPT = require("ai_commit_msg.prompts").DEFAULT_SYSTEM_PR
 ---@field max_tokens number|nil Maximum tokens in the response
 ---@field prompt string Prompt to send to the AI
 ---@field system_prompt string System prompt that defines the AI's role and behavior
----@field reasoning_effort string|nil Reasoning effort for models that support it ("minimal", "medium", "high")
+---@field reasoning_effort string|nil Reasoning effort for models that support it. Valid values:
+---  - "minimal": Fastest and cheapest. Least amount of "thinking". Best for simple, high-volume tasks like formatting or basic Q&A.
+---  - "low": Balance between speed and quality. Good for standard tasks like summarization.
+---  - "medium": Default-quality balance for creative and professional work.
+---  - "high": Most expensive and thorough; model performs deep, step-by-step reasoning (best for complex problems and debugging).
 ---@field pricing table|nil Pricing information for cost calculation. Supports:
 ---  - Flat table: { input_per_million, output_per_million } (backwards compatible)
 ---  - Map keyed by model: { ["model-name"] = { input_per_million, output_per_million }, default = { ... } }
@@ -107,6 +111,7 @@ M.default = {
       max_tokens = 10000,
       prompt = DEFAULT_PROMPT,
       system_prompt = DEFAULT_SYSTEM_PROMPT,
+      reasoning_effort = "minimal",
       pricing = {},
     },
   },

lua/ai_commit_msg/providers/copilot.lua

@@ -44,6 +44,11 @@ function M.call_api(config, diff, callback)
     payload_data.temperature = config.temperature
   end
 
+  -- Add reasoning effort for GitHub Copilot models if configured
+  if config.reasoning_effort then
+    payload_data.reasoning_effort = config.reasoning_effort
+  end
+
   local payload = vim.json.encode(payload_data)
 
   local curl_args = {

lua/ai_commit_msg/providers/openai.lua

@@ -53,7 +53,7 @@ function M.call_api(config, diff, callback)
     max_completion_tokens = config.max_tokens,
   }
 
-  -- Only add reasoning_effort for supported models
+  -- Only add reasoning (effort) for supported models
   if config.reasoning_effort and model_supports_reasoning_effort(config.model) then
     payload_data.reasoning_effort = config.reasoning_effort
   end