- Specialization
- Not classified
- Business
- Software Applications
- Market
- Software Development
- Ecosystem
- Agents
- Customers
Developersand 1 more customers valuesLess
- Enterprises
More details
- Organization type
- Commercial Company
- Ecosystem
- Agents
- Business
- Software Applications
- Specialization
- Not classified
- Market
- Software Development
- Customers
- Developers, Enterprises
- Delivery models
- Hosted Software
- Revenue models
- Subscription
- Listing status
- Not classified
- Latest equity round
- Not classified
- Reported funding amount
- Not classified
About
Fume builds an AI QA platform for software teams: customers describe tests in Loom videos, and Fume creates and maintains Playwright end-to-end browser tests. It differentiates itself as an autonomous QA department that designs, writes, and maintains resilient tests rather than merely providing a no-code test builder.
Market
Fume competes in AI-augmented software testing and end-to-end browser automation, positioning itself as an AI QA team that turns product videos and spoken requirements into maintained Playwright tests. It differentiates from broader cross-browser/device platforms and conventional test-automation tools through video-first authoring, portable Playwright code, automatic self-healing, and a browser-use agent that can recover from fragile scripted flows.
Fume targets growing software teams and larger engineering organizations that need autonomous end-to-end browser testing, particularly QA and engineering leaders responsible for regression coverage, test maintenance, and CI/CD quality. Its positioning is suitable for teams with web applications and existing Playwright, Selenium, or Cypress suites.
At a Glance
Problem
Software teams need reliable end-to-end browser tests to catch regressions in critical user flows, but conventional QA requires people to design, write, run, debug, and continually maintain those tests. That work is tedious, expensive, and competes with engineering time; it becomes especially painful as selectors, interfaces, UX flows, and application data change. The killer use case is turning a short recording of an important workflow—such as signup, login, checkout, or search—into durable tests quickly, so a team can cover its most important flows without building a large QA operation.
Product / Service
Fume positions itself as an AI QA team. A user records a screencast, often with Loom, while describing the intended behavior; Fume analyzes the video and application, extracts test cases, and generates Playwright end-to-end browser tests. Those tests can run on Fume's cloud runners or in the customer's CI/CD environment, with failures reported through shared channels and with an agent fallback that can complete runs when brittle scripted tests break. Fume then updates the test code to keep it stable as the product evolves, and customers retain ownership of the Playwright code. The service also supports migration of existing Playwright, Selenium, or Cypress tests.
Market
Fume competes in AI-augmented software testing, autonomous QA, and developer tools for end-to-end browser-test automation. Its differentiation is an autonomous, code-owning Playwright workflow rather than a conventional no-code test builder: Fume claims to design, generate, execute, and maintain tests, while adjacent alternatives include testRigor, mabl, BrowserStack, Testsigma, Katalon, and other AI-assisted testing platforms. The company is clearly beyond an idea stage: its site offers a $500-per-month Team plan, a 30-day free trial, free cloud test runners, and custom enterprise deployment; it also launched publicly on Product Hunt in 2025 and has reported $500,000 in seed funding. However, the available public evidence does not disclose a reliable revenue figure, so Fume should be described as a live, funded commercial product with revenue status undisclosed rather than as a company with verified scale or publicly reported revenue.
Founders & Leadership
Funding History
Y Combinator
Recent News
Skywork’s Fume deep dive reported that the platform integrates with developer tools including Slack, GitHub, GitLab, VSCode, Jira, and Linear. This is the clearest integration-related coverage found in the period.
Fume outlined its view that AI testing involves using large language models to write and maintain software QA tests. The article compares IDE assistance, AI-aided test builders, and fully agentic browser-testing systems.
Active Roles
0No active roles right now.
Get notified when they postBusiness Model
Fume sells its autonomous QA software as a recurring subscription, with a Team plan priced at $500 per month and custom pricing and deployment options for other customers.