Skip to content

Supported AI Providers

git-brief supports three major AI providers. Because the tool relies entirely on local Git metadata (commit messages, file paths) and not your source code, context windows are incredibly small.

A typical git brief execution consumes roughly ~500 input tokens and ~100 output tokens.

Google’s gemini-2.5-flash model is incredibly fast and completely free for this level of usage.

Anthropic’s claude-3-5-haiku-latest is their fastest, most cost-effective model, explicitly tuned for rapid text processing.

  • Cost per run: ~$0.00015 (Input) + ~$0.000125 (Output) = ~$0.00027
  • Cost for 1 year of daily standups: ~$0.07
  • Get an API Key: Anthropic Console

OpenAI’s gpt-4o-mini is their flagship small model.

  • Cost per run: ~$0.000075 (Input) + ~$0.00006 (Output) = ~$0.000135
  • Cost for 1 year of daily standups: ~$0.03
  • Get an API Key: OpenAI Platform