Find your next open-source project

Search by name, purpose, or keyword — or paste a GitHub repository URL
All skills

agent-platform-inference

Run a prompt against a Gemini or OpenMaaS model and return the model's answer with the exact project, region, model, and SDK used.

  • Part of google/skills
  • 9.6k installs
  • Machine Learning · Model serving

When to use it

You have a Google Cloud project and a specific model in mind but need working SDK code to call it, or you're hitting a 429, 400, or 404 error during inference.

For example

You say

Use agent-platform-inference to ask deepseek-ai/deepseek-v3.2-maas in project my-project, region global, via the OpenAI SDK: 'Explain vector databases in two sentences.'

You get
  • Python inference script
  • Model answer with parameters

How it works

  • Asks you to confirm project, region, model, SDK, and prompt before running any inference to avoid surprise cost.
  • Checks live regional availability for first-party Gemini models before writing code, instead of trusting docs.

Before installing

Needs Google Cloud Application Default Credentials (gcloud auth application-default login) and the aiplatform.googleapis.com API enabled, plus the google-genai, openai, or vertexai Python SDKs.

This skill includes 7 scripts your assistant may run on your computer.