Question: Copilot SDK Intended Scope & Cloud Deployment Support

[bluedog13]

Summary

I’m evaluating the GitHub Copilot SDK and would like clarification on its intended use cases, particularly around cloud-deployed and public-facing applications.

---

Current Understanding

From the documentation and experimentation, the Copilot SDK authenticates via:

• Local Copilot CLI login

This works well for:

• CLI tools used by individual developers
• Local applications tied to a signed-in GitHub user

---

Point of Confusion

It’s unclear how the Copilot SDK is intended to be used in cloud or server-side environments.

For example, in a public web application where:

• Any user can access the app (not limited to Copilot subscribers)
• Users do not authenticate with GitHub
• The backend handles AI requests on behalf of users

…the SDK does not appear to support this model. In such cases, developers would need to integrate directly with OpenAI, Anthropic, or Gemini APIs instead of using the Copilot SDK.

---

Questions

• Is the Copilot SDK intentionally scoped to local and developer tooling use cases?
• Are there plans to support cloud deployment scenarios (e.g., service accounts, server-side credentials, or API keys)?
• For public-facing applications, what is the recommended approach for achieving multi-model support similar to Copilot’s abstraction layer?

---

Why This Matters

One of the SDK’s main value propositions is a unified API across multiple AI providers (GPT, Claude, Gemini). If the SDK is limited to local or authenticated developer environments, teams building cloud applications must:

• Integrate with each AI provider separately
• Manage multiple API keys and billing accounts
• Miss out on Copilot’s unified model selection experience

Clarifying the intended scope and roadmap would help developers choose the right tool early and avoid incorrect architectural assumptions.

Thanks

[github-actions]

Thanks for opening this issue, @bluedog13! This is a well-articulated question about the SDK's intended scope and deployment scenarios.

I've labeled this as question and documentation since the answers would help clarify the SDK's architecture and use cases for others.

The questions you've raised about cloud deployment, service accounts, and server-side credentials are important for teams evaluating the SDK. A maintainer will review and provide guidance on:

• The SDK's current scope (local vs. cloud environments)
• Any planned support for server-side deployment scenarios
• Recommended approaches for public-facing applications

Thanks for helping improve the documentation!

AI generated by Issue Triage Agent