GitHub has unveiled a new feature allowing developers to run multiple AI-assisted tasks simultaneously within Copilot, enhancing productivity and project management through isolated, parallel workstreams.
GitHub has added parallel agent sessions to its Copilot app, allowing developers to run several AI-assisted tasks at the same time without the sessions interfering with one another. According to GitHub’s documentation, each session is isolated in its own Git worktree, which means different jobs can proceed in parallel while keeping code changes separate and easier to review.
The new capability is aimed at developers working on multi-part projects. Instead of waiting for one task to finish before starting another, users can assign Copilot to build a feature, carry out an accessibility review and run test suites concurrently. GitHub presents the approach as a way to reduce context switching and keep the developer focused on higher-level decisions rather than manual coordination.
GitHub’s own guidance says agent sessions can be started in different modes, including Interactive, Plan and Autopilot, with users able to choose a model and reasoning effort before submitting a prompt. The company also says the Copilot app can be used across paid plans, while its cloud agents can be launched from GitHub, GitHub Mobile, code editors and the GitHub CLI, as well as through the REST API.
The company has also built management tools around the feature. An agents panel allows users to monitor activity, review logs, steer a session with follow-up prompts, stop or archive work and continue a task in Copilot Chat. That matters because the practical value of parallel agents depends not just on starting them, but on keeping visibility across multiple active runs.
Disclaimer: This content is intended for informational purposes only. Readers are advised to exercise their own judgement, conduct due diligence, or consult a qualified expert before acting on any information provided.
