Skill

Google Gemini

Google's Gemini models (and on-device Gemma) — the default provider in several of my open-source agent projects.

  • #ai
  • #providers

Gemini is Google's family of multimodal LLMs, available through the Gemini API; Gemma is the related family of open models that can run on-device.

How I use it

  • Akashi — akashi_google is a provider adapter (structured output, tool choice, embeddings) that powers the quick-start streaming agent. akashi_gemma runs Gemma on-device in Flutter, proving the model contract is not tied to HTTP.
  • langgraph-langserve-starter — Gemini Flash is the default model.
  • local-rag-chatbot — Gemini is one of the selectable cloud providers.
  • Somnio AI Tools — the CLI installs agent skills across AI coding CLIs, including Gemini's.

Why it matters

Using Gemini as a default in open-source examples keeps the barrier to entry low for anyone cloning the repos, while the provider-neutral design means switching to OpenAI or Anthropic Claude is a configuration change rather than a code change.

Related