AI test authoring and external agents
MagicPod Autopilot creates and edits tests from Japanese text instructions. The MagicPod MCP server also lets Claude, Codex, Cursor, and Devin operate MagicPod through natural language.
MagicPodとは？
MagicPod is a cloud service for automating browser and mobile app tests. Teams build test scenarios by selecting AI-detected screen elements in a no-code editor or issuing instructions to an AI agent. Its readable test scripts let people without programming knowledge participate in test creation and maintenance.
The platform combines test authoring, execution, result reports, and maintenance tools. Teams can use MagicPod's cloud environments, local devices, or supported external device clouds, with integrations for existing development workflows.
MagicPodでできること
Screen checks and automatic repair
AI assertions evaluate screen content against expectations written in natural language. During batch runs, automatic repair attempts to fix locator definitions and commands after a test fails.
Browser and mobile coverage
Run tests across browsers and iOS or Android devices, including WebView and Flutter apps. Commands cover file uploads, browser storage, gestures, device rotation, and other platform-specific actions.
Reusable test logic
Data-driven tests repeat scenarios with different input values, including CSV data. Variables, conditional branches, loops, and explicit waits control execution without writing code.
Parallel execution and delivery integrations
Execute tests in parallel using cloud environments. Integrations include GitHub Actions, Jenkins, CircleCI, and Bitrise, plus BrowserStack, Sauce Labs, and Remote TestKit for external test infrastructure.
Failure analysis and project health
AI failure analysis investigates unsuccessful runs. Result reports show execution outcomes, while project health scores identify areas for improvement and suggest changes.
利用シーン
“Recurring regression checks”
QA teams can move repeated release checks into scheduled or CI-triggered runs, using reports to investigate failures before shipping changes.
“Device and browser compatibility testing”
Validate the same customer journey across browser configurations, mobile OS versions, and device models rather than repeating each check manually.
“Manage browser and mobile QA together”
Teams maintaining both interfaces can manage and batch-run their browser and mobile test cases within one project using mixed-platform testing. Each testing type still requires its own contract.
“Establish a shared QA practice”
Optional onboarding support covers test reviews, failure diagnosis, training, and operating rules so the team can learn to maintain its own automation. Scope and fees are discussed separately.
よくある質問
Can I try MagicPod before subscribing?
A two-week free trial lets you create and execute tests without registering a credit card. Continued use is through a paid plan.
How much does Standard cost?
The English Standard plan costs $320/month with an annual contract or $400/month with a monthly contract, excluding tax, for one testing plan. Japanese billing lists ¥39,800 and ¥54,800 respectively. Enterprise pricing is negotiated.
What limits apply to Standard?
It starts with three projects, 100 test cases, 100 visual difference check steps, and 10 branches, with two concurrent cloud environments each for creation and batch runs. Users and executions are unlimited. Test-case capacity is shared across projects, includes trashed cases, and counts cases over 300 steps as multiple units. Paid add-ons are available.
Does one contract cover browser and mobile testing?
No. Browser and mobile app testing require separate contracts, including when their cases are managed in one mixed-platform project. Both contracts must use the same contract period and payment method. External device-cloud services also need separate service contracts.
What does Enterprise add?
It removes limits on test cases, visual difference check steps, and branches, and adds SAML SSO, IP restrictions, domain control, secure tunneling, contract customization, and premium support. Test results are retained for two years rather than Standard's one year. Project counts and cloud concurrency are arranged with MagicPod.
How are test assets and AI inputs handled?
Test assets can include application binaries, UI images, logs, and prompts. MagicPod uses cloud infrastructure and may send information to providers such as Anthropic and OpenAI for AI features or support. Its terms state that user information is not used to train its own or external generative AI models.
クイック情報
- Product category
- AI-assisted test automation
- Test targets
- Web browsers and iOS/Android apps
- Customer support languages
- English and Japanese
- External cloud services
- Separate service contracts required
- Website
- https://magicpod.com
- Pricing model
- Paid testing plans with a two-week trial
MagicPod のトラフィック分析
トラフィックデータは参考情報としてご利用ください。
- ドメイン評価
- 53
MagicPodの代替品
Applitools
applitools.com
Applitoolsは、製品ライフサイクル全体にわたり、Web、モバイル、デスクトップ、コンポーネント、ドキュメントのエクスペリエンスを検証するAIテスト自動化プラットフォームです。
Katalon
katalon.com
AI-assisted test automation and test management for software QA teams
Rainforest QA
rainforestqa.com
SaaSチームとCI/CDに対応するノーコードAIテスト自動化
mabl
mabl.com
Web、モバイル、AIアプリ、API向けのエージェント型テスト自動化
Testim
testim.io
Record application flows, add checks and run end-to-end tests. Testim combines AI-assisted element locators with visual test editing and team workflows for web, mobile and Salesforce testing.
ACCELQ
accelq.com
Web、API、モバイル、デスクトップ、バックエンド、パッケージアプリに対応するクラウド型AIノーコードテスト自動化。