CircleCI

Cloud and self-hosted CI/CD with a strong reputation for fast pipelines and config-as-code. Native AI features: AI Test Insights for flaky test detection, ML-driven test splitting, and pipeline anomaly detection. Resource classes from small to GPU. Integrates with every major SCM and cloud, with first-class macOS for iOS builds.

Deploy Secure

Visit website

Last reviewedMay 15, 2026

How we review

Overview

CircleCI is a container-native CI platform. Pipelines are defined in .circleci/config.yml using a YAML model that distinguishes jobs (units of work), steps (commands within a job), and executors (compute environment: Docker, machine, macOS, Windows, GPU). Jobs within a workflow run in parallel or in dependency order.

Orbs are reusable, versioned packages of CircleCI configuration that encapsulate jobs, commands, and executors, distributed via the Orb Registry. Docker Layer Caching stores intermediate image layers between runs. Test splitting distributes test files across parallel containers based on historical run-time data so all containers complete at approximately the same time.

Resource classes configure compute per job (CPU, memory, OS), enabling jobs with different compute requirements to run on appropriately sized machines within the same workflow. A self-hosted runner option and an on-premises server product cover isolated environments. Pricing is credit-based; credits are consumed at different rates per resource class. Free tier: 6,000 credits/month on Linux.

Key Features

Orbs: reusable, versioned, composable config packages that encapsulate jobs, commands, and executors — stronger typing and composability than GitHub Actions' marketplace model for complex pipeline logic
Docker Layer Caching (DLC): caches intermediate Docker image layers between pipeline runs, reducing build times on image-heavy workflows without storing full images in an external registry
Test splitting: automatically distributes test files across parallel containers based on historical timing data, balancing workload so all containers finish at approximately the same time
Resource classes: configurable compute per job (Docker, machine, macOS, Windows, GPU) with memory and CPU sizing — jobs with different compute needs run on appropriate hardware within the same workflow
Pipeline config splitting for monorepos: dynamic config generation allows large repositories to only run the pipelines relevant to changed paths, avoiding full-suite runs on unrelated changes
Insights: per-workflow and per-job analytics showing build duration trends, flaky test rates, and credit consumption — surfaces which jobs are the slowest and most expensive across the pipeline

Integrations

8 total

scm

GitHubGitLabBitbucket

ci / cd

Terraform

cloud

AWSGCP

orchestration

Kubernetes

messaging

Slack

Pricing

No pricing data on file.

Compare with similar tools

1 comparison

CircleCI vs GitHub Actions

Compliance

2 / 6 attested

SOC 2

HIPAA

GDPR

FedRAMP

PCI DSS

ISO 27001

Freshness

Heartbeatverified · 2d ago

Listings are re-verified on a recurring sweep. If the website 200s and pricing hasn't shifted since the last check, the heartbeat stays green.

UpdatedOct 22, 2026

What does AI Enhanced mean?

AI maturity is an independent editorial rating assigned by infraplz.dev — not a vendor self-claim. We assess how meaningfully each tool actually uses AI in production workflows, cutting through marketing copy.

AI NativeAI is the core product — the tool wouldn't exist without it.
AI EnhancedAI is a first-class feature that materially changes the workflow.
AI AdjacentAI features exist but are superficial or bolt-on additions.
AI MinimalNo meaningful AI features; listed as a critical platform engineering tool.

Spot something wrong on this page? Report a correction →