See how the GitHub Copilot SDK enables agentic workflows directly inside your applications.

The GitHub Blog provides updates, announcements, and insights from the world's leading software development platform, covering topics such as new features, community highlights, and industry trends. Developers can learn about GitHub's latest developments, best practices for collaboration, and tips for maximizing productivity on the platform.

GitHub Blog

GitHub has released the Copilot SDK, which lets developers embed agentic execution capabilities directly into their own applications. Rather than relying on text-in/text-out AI interactions, the SDK exposes the same planning and execution engine that powers GitHub Copilot CLI as a programmable layer. Three key patterns are highlighted: delegating multi-step work to agents instead of brittle scripts, grounding execution in structured runtime context via Model Context Protocol (MCP), and embedding agentic execution outside the IDE into desktop apps, SaaS platforms, background services, and event-driven systems.

The era of “AI as text” is over. Execution is the new interface.

Pattern #1: Delegate multi-step work to agents

Pattern #2: Ground execution in structured runtime context

Pattern #3: Embed execution outside the IDE

“delegating multi-step work to agents instead of brittle scripts” since when is an AI agent more robust than a script? Not to shit on AI but at least a script will have a deterministic output

When I say delete them all, it says there aren’t enough tokens… - -’

Well said. AI gets far more useful when it moves beyond responses and starts operating inside real workflows.