FumeDev
Automates coding, integrates seamlessly, enhances developer productivity.. [Contact for Pricing]
Last verified:
What is FumeDev?
Fume is an AI-powered QA tool that acts as your autonomous AI QA team, creating and maintaining Playwright end-to-end browser tests for web applications. Users simply record a short Loom-style screencast showing their application workflow while verbally describing what should be tested, and Fume automatically extracts test cases and generates complete Playwright test suites in seconds.
Key features include self-healing tests that automatically update when UI selectors change, AI agent fallbacks that pick up where broken scripts left off, free cloud test runners that execute tests twice daily, full Playwright code ownership (you own the code even if you cancel), Slack notifications for test results, CI/CD integration via API for GitHub Actions/GitLab/CircleCI, and the ability to migrate existing Playwright/Selenium/Cypress tests with instant AI maintenance.
Fume is designed for software development teams, startups, product managers, and QA teams who want to eliminate manual test writing and maintenance. It's especially valuable for teams without dedicated QA resources, growing teams needing autonomous testing, and organizations wanting to cover 80% of important user flows in under 10 minutes without code lock-in.
FumeDev pricing
Pricing model: Freemium
Team Plan: $500/month with 30-day free trial, includes 10 seats per year (can add more seats later), perfect for growing teams needing autonomous QA. Enterprise: Contact sales for custom pricing and deployment options, custom solutions for large organizations with volume discounts typical above 50 seats. Free test runners on Fume Cloud are included. Each account is a team account with a single seat by default. No free tier - only the 30-day free trial before paid subscription.
FumeDev pros
- Generates Playwright tests fromLoom video walkthroughs in seconds
- Self-healing tests automatically update when UI changes
- AI agent fallbacks complete runs when scripts fail
- Free cloud test runners with twice-daily execution
- 100% Playwright code ownership even after cancellation
- No setup required - tests run immediately after generation
- Slack integration for real-time test failure notifications
- Covers 80% of important user flows in under 10 minutes
- Migrates existing Playwright/Selenium/Cypress tests instantly
- CI/CD integration via API for GitHub Actions and GitLab
- Runs tests locally or in your own CI machines
- Handles dynamic app data during runtime execution
- Replaces entire QA team operations for browser testing
- No app installation needed for test execution
- 30-day free trial available for team plan
FumeDev cons
- No free tier - only 30-day free trial then $500/month
- Team plan at $500/month may be expensive for small teams
- Requires Loom or screen recording quality for best results
- GitHub integration included by default, other integrations need contact
- Studio setup takes a few minutes to boot up
- Only supports web applications, not mobile apps
- Video clarity directly impacts test generation accuracy
- Enterprise pricing requires contacting sales, not transparent
Frequently asked questions about FumeDev
What is Fume and how does it work?
Fume is your AI QA team that creates and maintains Playwright end-to-end browser tests. You record a short Loom video showing your application flow while describing what to test. Fume analyzes the video and your app to automatically generate Playwright tests, runs them in the cloud or your CI, and self-heals tests when UI changes occur.
Do I own the Playwright code Fume generates?
Yes, you always own the output Playwright code even if you cancel your subscription. Fume tests are 100% in Playwright, giving you full codebase ownership. You can run tests locally or anywhere, even if you stop paying.
What happens when a test fails due to UI changes?
Fume has AI agent fallbacks that pick up exactly where the script left off using a browser-use agent. The flaky script is then self-healed in the background by updating selectors and flows automatically, so future runs pass without manual intervention.
Where can I run Fume tests?
You can run tests on Fume Cloud using their free test runners, or integrate into your own CI/CD pipeline via API. Fume supports GitHub Actions, GitLab CI, CircleCI, and custom CI systems. Tests execute twice daily on Fume Cloud by default.
Can I migrate my existing tests to Fume?
Yes, Fume supports migrating existing Playwright, Selenium, or Cypress tests. When you migrate, your tests instantly get native AI fallbacks and auto-maintenance capabilities without rewriting them from scratch.
What video tools work for recording test flows?
Fume works with any screencast tool including Loom, Mac Screen Recorder, Tella, and similar tools. Record 30-90 seconds showing how to navigate pages, inputs/interactions, and expected success outcomes while describing what should be tested.
How many seats are included in the Team Plan?
The Team Plan at $500/month includes 10 seats per year by default. Each account starts as a team account with a single seat, but you can add more seats later. The number of engineers in your team is used to calculate seats allocated during the free trial.
What integrations are available by default?
Team accounts have access to Github integration by default. If you need GitLab or other integrations, you must contact Fume at [email protected] or book a time to discuss your use case. API keys are found in your Fume dashboard at app.fumedev.com.
How is Fume different from no-code test builders?
Fume is not a no-code test builder - it's a fully autonomous QA department. Unlike no-code builders where humans design tests and tests break with changing UI, Fume's AI designs resilient tests, handles major UX changes agenticially, handles dynamic app data at runtime, and provides 100% Playwright code ownership without vendor lock-in.
How do I get started with Fume?
Sign up for a team account at app.fumedev.com/signup, add your company name and work email. Connect your code repository via Github integration, specify the number of engineers on your team, create your first studio with a name, and then record your first Loom video to generate tests. The studio takes a few minutes to boot up before it's ready.