Provider Configuration

gcop-rs supports multiple LLM providers. You can use built-in providers or add custom ones.

Built-in Providers

Claude (Anthropic)

[llm.providers.claude]
api_key = "sk-ant-your-key"
model = "claude-sonnet-4-5-20250929"
temperature = 0.3
max_tokens = 2000

Get API Key: https://console.anthropic.com/

Available Models:

• claude-sonnet-4-5-20250929 (recommended)
• claude-opus-4-5-20251101 (most powerful)
• claude-3-5-sonnet-20241022 (older version)

OpenAI

[llm.providers.openai]
api_key = "sk-your-openai-key"
model = "gpt-4-turbo"
temperature = 0.3

Get API Key: https://platform.openai.com/

Available Models:

• gpt-4-turbo
• gpt-4
• gpt-3.5-turbo

Ollama (Local)

[llm.providers.ollama]
endpoint = "http://localhost:11434/api/generate"
model = "codellama:13b"

Setup:

# Install Ollama
curl https://ollama.ai/install.sh | sh

# Pull a model
ollama pull codellama:13b

# Start server
ollama serve

Available Models: Any model available in Ollama (codellama, llama2, mistral, etc.)

Gemini (Google)

[llm.providers.gemini]
api_key = "AIza-your-gemini-key"
model = "gemini-3-flash-preview"
temperature = 0.3

Get API Key: https://ai.google.dev/

Available Models:

• gemini-3-flash-preview (recommended default)
• gemini-2.5-flash
• gemini-2.5-pro

Custom Providers

You can add OpenAI-, Claude-, or Gemini-compatible APIs using the api_style parameter.

DeepSeek

[llm.providers.deepseek]
api_style = "openai"
api_key = "sk-your-deepseek-key"
endpoint = "https://api.deepseek.com/v1/chat/completions"
model = "deepseek-chat"
temperature = 0.3

Get API Key: https://platform.deepseek.com/

Qwen (通义千问)

[llm.providers.qwen]
api_style = "openai"
api_key = "sk-your-qwen-key"
endpoint = "https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions"
model = "qwen-max"

Claude Proxy/Mirror

[llm.providers.claude-code-hub]
api_style = "claude"
api_key = "your-key"
endpoint = "https://your-claude-code-hub.com/v1/messages"
model = "claude-sonnet-4-5-20250929"

Custom OpenAI Compatible Service

[llm.providers.my-llm]
api_style = "openai"
api_key = "your-key"
endpoint = "https://api.example.com/v1/chat/completions"
model = "custom-model"

API Style Parameter

The api_style parameter determines which API implementation to use:

Value	Description	Compatible Services
"openai"	OpenAI Chat Completions API	OpenAI, DeepSeek, Qwen, most custom services
"claude"	Anthropic Messages API	Claude, Claude proxies/mirrors
"ollama"	Ollama Generate API	Local Ollama only
"gemini"	Google Gemini GenerateContent API	Gemini and Gemini-compatible endpoints

If api_style is not specified, it defaults to the provider name (for backward compatibility with built-in providers).

Switching Providers

Using Command-Line

# Use different provider for one command
gcop-rs --provider openai commit
gcop-rs --provider deepseek review changes

Changing Default

Edit your platform-specific config file (see Configuration Guide):

[llm]
default_provider = "deepseek"  # Change this

API Key Management

Config File

Provider api_key is configured in config.toml:

[llm.providers.claude]
api_key = "sk-ant-..."

CI Mode Environment Variables

In CI mode (CI=1), use environment variables instead of config file:

• GCOP_CI_PROVIDER - Provider type: claude, openai, ollama, or gemini
• GCOP_CI_API_KEY - API key
• GCOP_CI_MODEL (optional, has defaults)
• GCOP_CI_ENDPOINT (optional)

See Also

• Configuration Reference - All configuration options
• Provider Health Checks - How gcop-rs config validate works
• Custom Prompts - Customize AI behavior
• Troubleshooting - Provider connection issues