[FEATURE] Native CLI Support for OpenAI, Anthropic, and Gemini APIs

[nirvik34]

Gitbun currently relies heavily on local Ollama instances for the CLI. We need to expand src/llm/ to support remote APIs. The user should be able to set backend: "openai" in their config and provide an LLM_API_KEY in their environment variables. This requires building a modular provider interface.

[saksenaaishwarya]

I'd like to work on this under GSSoC'26.

My plan:

• Create a src/llm/providers/ directory with a base LLMProvider interface/class
• Implement three concrete providers: OpenAIProvider, AnthropicProvider, and GeminiProvider, each reading their respective API key from env (OPENAI_API_KEY, ANTHROPIC_API_KEY, GOOGLE_API_KEY)
• Add a backend field to the Gitbun config file (e.g., gitbun.config.json) that accepts "openai", "anthropic", "gemini", or "ollama" (default)
• A ProviderFactory that reads the config and instantiates the right provider at runtime
• Graceful error messages when the API key is missing or the provider name is invalid
• Update CLI --backend flag to allow inline override
• Keep Ollama as the fallback default so existing users aren't affected

Please assign this to me if it's still available!