Cirrus CI is shutting down: upgrade to a scalable, AI-ready alternative

On April 7, the Cirrus Labs team announced they are joining OpenAI. As a result, Cirrus CI will stop running jobs on June 1, 2026.

Cirrus CI was built for engineering teams who demand high-performance hardware and clean configuration. As you look for a new home, you shouldn’t have to compromise on those standards. CircleCI matches the infrastructure and config model Cirrus teams depend on, and adds the orchestration, reusable config, and AI-native tooling to take your pipelines further.

Below, we’ll cover what maps directly from Cirrus CI to CircleCI, where CircleCI goes further, and how to start migrating before the June 1 deadline.

Start your migration · Talk to a migration expert

What carries over from Cirrus CI

Config-as-code. Your pipeline is defined in portable, version-controlled YAML files stored in your repo by default. You can also store config in a separate repository at any path, which enables centralized config management across projects.

Pay for what you use. CircleCI bills compute down to the second, so you only pay for the resources your team uses. Other providers, like GitHub Actions, round up to the nearest minute, which adds significant cost at scale.

Multi-platform execution. Docker, Linux (x86 and Arm), macOS, Windows, and GPU execution environments. Resource classes expose real CPU and memory allocations down to the job level, so you size compute to the workload.

Docker-native workflows. Define your environment with any image, run service containers alongside it (a direct replacement for additional_containers), and use Docker Layer Caching to speed up image builds. CircleCI’s convenience images (cimg/*) are purpose-built for CI and updated regularly.

Mobile CI. If you’re building iOS or Android on Cirrus, CircleCI supports the same workflows: macOS executors, Docker for Android, and pre-built Fastlane integration for code signing and app store deployment.

Bring your own compute. Self-hosted runners let you run jobs on your own infrastructure (Kubernetes clusters, bare metal, VMs, or cloud instances) while CircleCI handles orchestration, logging, and reporting. If you ran Cirrus tasks on your own GCP, AWS, or Azure infrastructure, this is the direct equivalent.

Open source support. Cirrus CI’s generous free tier made it a home for projects like PostgreSQL, Bitcoin Core, Podman, and Flutter. CircleCI’s open source program provides up to 400,000 credits per month for Linux, Arm, and Docker builds, plus 30,000 for macOS on public repos. That’s over 13x the standard free tier, with access to the same features available on paid plans.

What CircleCI adds

VCS flexibility

Cirrus CI supported GitHub only. CircleCI integrates natively with GitHub, GitLab, and Bitbucket.

macOS on M4 Pro

Cirrus ran macOS builds on M1 through Tart. CircleCI’s managed fleet runs on Apple M4 Pro by default: more CPU cores, better GPU performance, and up to 56GB of unified memory. We support Xcode 13.4.1 through the latest 26.x releases, with m4pro.medium and m4pro.large resource classes. For teams that need their own Apple hardware, self-hosted runners support macOS natively.

Common mobile CI tasks (building, testing, code signing, app store deployment) are pre-packaged in the Fastlane orb, so you’re not writing that plumbing from scratch.

Faster testing and test insights

CircleCI gives you two ways to cut test execution time. Intelligent test splitting distributes suites across parallel containers using historical timing data, so each container finishes at roughly the same time. Smarter Testing goes further by analyzing code diffs, dependency graphs, and historical test behavior to run only the tests impacted by a change, then dynamically distributing them across parallel workers. Early users are seeing up to 4x faster feedback cycles.

On the visibility side, store_test_results feeds the Insights dashboard for flaky test detection, timing trends, and optimization over time. Cirrus CI had no equivalent for any of this.

Configuration and orchestration at scale

CircleCI gives you the building blocks to standardize configuration, share it across teams, generate pipelines dynamically, and control how code moves to production.

• Orbs are shareable, parameterized packages of CI/CD configuration: pre-built jobs, commands, and executors for AWS, GCP, Kubernetes, Terraform, Slack, security scanning, and more. Thousands are available in the public registry, and you can create private orbs for your organization.
• Dynamic configuration goes beyond Cirrus CI’s matrix builds and conditional task execution. You can generate or modify your entire pipeline graph at runtime based on which files changed, which branch triggered the build, or any other condition.
• Platform Team Toolkit lets platform teams manage pipelines from centralized config templates with overrides, allowlists, a Terraform provider, and project admin APIs. Cirrus had no equivalent.
• Release orchestration adds deployment timeline views, environment hierarchies with enforced promotion paths, and rollback pipelines for visibility and control over how code moves from staging to production.

AI-native tooling

The Cirrus Labs team cited the shift to agentic engineering as their reason for joining OpenAI. CircleCI is already building for that shift inside the CI/CD layer.

• The MCP Server connects your AI coding assistant (Cursor, Claude Code, Windsurf, Copilot, or any MCP-compatible client) directly to your pipelines. Your assistant can retrieve logs from failed builds, diagnose issues, and suggest fixes through natural language, without leaving your editor.

• Chunk is an autonomous CI/CD agent with specialized skills for fixing flaky tests, writing new tests, diagnosing bugs, refactoring code, improving documentation, and optimizing build configs. Chat with Chunk in natural language, assign scheduled tasks that open PRs autonomously, or use contextual “fix” buttons on failed builds in the UI.

• The Chunk CLI brings Chunk to your local workflow. chunk hook wires test, lint, and AI code review into your coding agent’s lifecycle (Claude Code, Cursor, Copilot), and chunk build-prompt generates review context from your team’s actual PR history.

Security and compliance

CircleCI is SOC 2 Type II compliant and was the first CI/CD provider to achieve FedRAMP authorization.

For security-conscious orgs, config policies let platform teams enforce checks before pipelines run. Contexts provide scoped secret management with OIDC support for credential-free cloud authentication.

CircleCI is also a trusted publisher for npm trusted publishing, so maintainers can publish packages directly from CI without storing credentials.

How to migrate

Migrating CI providers is never trivial, but moving from Cirrus CI to CircleCI is one of the easier transitions you can make. CircleCI uses a similar YAML-based config syntax, and the core concepts you already work with (executors, caching, artifacts, dependency ordering) all have direct equivalents. Most teams can get a working pipeline in an afternoon and optimize from there.

If you want hands-on help, CircleCI’s technical services team can assist with migration planning and pipeline architecture.

Here’s how the key concepts map:

Cirrus CI	CircleCI
task	job
container	docker executor
additional_containers	Secondary images in docker executor
machine (Linux VM)	machine executor
macos_instance (M1/Tart)	macos executor (M4 Pro)
script instruction	run step
cache instruction	save_cache / restore_cache
artifacts instruction	store_artifacts
matrix modifier	matrix in job parameters
depends_on	requires in workflows
Encrypted variables	Contexts or project env vars

For full details on executors, caching, workflows, and every other configuration option, see the configuration reference.

Start building on CircleCI

CircleCI gives you the infrastructure Cirrus teams expect, with the orchestration, test intelligence, and AI-native tooling to keep pushing forward. Sign up free to get started, no credit card required.

Open source maintainers can apply for expanded credits, and early-stage companies may qualify for up to $20,000 in free compute credits through the startup program. If you want hands-on help with migration, CircleCI’s technical services team is here.

Exploring other options? See how CircleCI compares.

Start your migration · Talk to a migration expert