Which Google API key to use? Vertex AI API or an API key created in Goggle AI Studio?

[Jmcassociates]

Question

I use the Vetex AI API key in GCP for mciroservices I have on a project I'm building for the AI portion. Can I gen another API key with Vertex and restrict it to opencode? What's the bsst method to use 2.5 pro?

[github-actions]

This issue might be a duplicate of existing issues. Please check:

• Google VERTEX AI bug with MCP servers #1725: Google VERTEX AI bug with MCP servers - User had issues with Vertex AI API and asked about using Gemini Pro 2.5 with Vertex AI
• Error: Could not load the default credentials. Browse to https://cloud.google.com/docs/authentication/getting-started for more information. #553: Error: Could not load the default credentials - Discusses authentication issues when trying to use Vertex API
• Supporting Vertex-based models using google auth #272: Supporting Vertex-based models using google auth - Detailed discussion about setting up Vertex AI with Google authentication vs regular API keys
• Allow using Gemini CLI for free Gemini requests #402: Allow using Gemini CLI for free Gemini requests - Discusses using Google AI Studio vs other Google APIs for Gemini access
• Any gemini vertex support? i can't see it onthe list of models.. #1204: Any gemini vertex support? - User asking about Vertex support for Gemini models

These issues cover similar topics about Google API authentication, Vertex AI setup, and choosing between different Google AI services. The discussion in #272 particularly covers the exact question about Vertex AI vs Google AI Studio API keys.

Feel free to ignore if none of these address your specific case.

[rekram1-node]

I think we discussed this in discord right