AI Provider Setup

Zgit supports multiple AI providers for commit messages, PR generation, and other smart features.

Supported Providers

Provider	Models	Auth
OpenAI	GPT-4o, GPT-4o-mini, o1, o3	Bearer token
Anthropic	Claude 3.5 Sonnet, Claude 3 Opus	x-api-key header
Gemini	Gemini 1.5 Pro, Gemini 1.5 Flash	API key (query param)
DeepSeek	DeepSeek-V2, DeepSeek-Coder	Bearer token
Groq	Llama 3, Mixtral, Gemma	Bearer token
OpenRouter	200+ models via unified API	Bearer token
Ollama	Any local model	No auth (localhost)

Adding a Provider

1. Open Settings → AI
2. Click Add Provider
3. Select the provider from the dropdown
4. Enter your API key
5. Choose a model (fetch dynamically or type manually)
6. Toggle Set as Active

Encrypted Key Storage

API keys are encrypted with AES-256-GCM before being saved to disk.

• Encryption key auto-generated at ~/.config/zgit/.ai_key_encryption (0600 permissions)
• Plaintext keys never sent to the frontend
• UI shows only masked previews: sk-...abcd
• Keys persist per-provider; delete via Clear button

Dynamic Model Fetching

Each provider’s available models can be fetched live:

Provider	Endpoint
OpenAI	GET https://api.openai.com/v1/models
Groq	GET https://api.groq.com/openai/v1/models
OpenRouter	GET https://openrouter.ai/api/v1/models
Ollama	GET http://localhost:11434/api/tags

Fetched models appear as clickable chips. Select one or type a custom model ID.

Configuration File

Keys are stored in ~/.config/zgit/config.yaml under:

ai:
  active_provider: openai
  providers:
    openai:
      api_key: <encrypted>
      model: gpt-4o
      endpoint: https://api.openai.com/v1

Use zgit config set-ai --provider openai --model gpt-4o from the CLI.