GitHub has introduced the GitHub Copilot SDK in technical preview, allowing developers to embed Copilot’s agentic capabilities directly into their own applications.
The SDK exposes the same execution loop used by GitHub Copilot CLI, including planning, tool invocation, file editing, and command execution. According to GitHub, this is intended to reduce the complexity of building agent-based systems from scratch.
“Building agentic workflows from scratch is hard,” said the chief product officer, Mario Rodriguez, in a blog post. “Even before you reach your actual product logic, you’ve already built a small platform.”
GitHub said the Copilot SDK provides programmatic access to Copilot’s production-tested agent loop, removing the need for developers to design their own planners and runtimes. The SDK supports multiple AI models, custom tool definitions, MCP server integration, GitHub authentication, and real-time streaming.
The technical preview initially supports Node.js, Python, Go, and .NET. Developers can use an existing GitHub Copilot subscription or supply their own API key. The open repository includes setup instructions, starter examples, and SDK references for each language.
GitHub recommends starting with a single task, such as updating files or running commands, and allowing Copilot to plan and execute steps while the host application provides tools and constraints. In an example shared by GitHub, it was revealed that developers can create a Copilot client, start a session using a specified model, and send prompts programmatically.
The company said the SDK builds directly on the capabilities of Copilot CLI, which already allows users to plan projects, modify files, run commands, and delegate tasks without leaving the terminal. Recent updates to Copilot CLI include persistent memory, multi-step workflows, full MCP support, and asynchronous task delegation.
“The SDK takes the agentic power of Copilot CLI and makes it available in your favourite programming language,” Rodriguez wrote. “This makes it possible to integrate Copilot into any environment.”
Internal GitHub teams have used the SDK to build tools such as YouTube chapter generators, summarisation tools, custom agent interfaces, and speech-to-command workflows, according to the company.
GitHub positioned the Copilot SDK as an execution layer, with GitHub managing authentication, model access, and session handling, while developers control how those components are used within their applications.
