TesterArmy (YC 26) runs agents that test your web and mobile app like real users on every pull request. Describe a flow in plain language and get passed, failed or blocked with a recording and a transcript. Pass it to your coding agent.
A startup from San Fransisco, the United States that is founded by Szymon Rybczak, Oskar Kwasniewski.
This page is designed to help you find out whether TesterArmy is good and if it is the right choice for you.
Writing end-to-end tests takes longer than the feature, the scripts break on every UI change, and when they pass they only prove what you told them to check.
So instead of scripts, an agent uses the app the way a user would, on every pull request. Open a PR and it runs the flows you care about as a check, or point it at production on a schedule. It works on web, iOS and Android, including Expo and EAS builds, and it handles login, OAuth, one-time passcodes and test accounts.
You describe each flow in plain language, something like “log in as the test user, open the API keys page, create a key, check it shows up”. The agent opens a real browser or device, works out what to click from what it sees, and remembers what it learned about your app for the next run. There is a CLI and an MCP server, so Claude Code, Codex or Cursor can write and run these tests for you.
Every run ends with one of three verdicts: passed, failed, or blocked. Blocked means the agent could not get through for a reason outside your product, like a third-party outage, so you do not chase it as a bug. Each run keeps a recording, screenshots, the browser console and network logs, and a step-by-step transcript of what the agent did and why.
When something fails, paste the report into your coding agent and let it fix the bug, or send it to Linear as an issue. Results also go to GitHub, Slack, Discord or a webhook.
Listed in
Resend, bolt.new, Nando's, Novu, Rork, August
The agent tests the app the way a person does. It looks at the screen, decides what to click, and follows a flow you wrote in plain language, so there are no selectors or scripts to maintain when the UI changes. Every run ends in one of three verdicts, passed, failed or blocked, and blocked is kept separate so a third-party outage never shows up as a bug in your product. It is built to be driven by other agents too: the CLI and the MCP server let Claude Code, Codex or Cursor create tests, run them and read the transcript, and the failure report is written to be pasted straight back into a coding agent. Web, iOS and Android, including Expo and EAS builds, in one tool.
If you are on Playwright, Cypress or Selenium, you know the cost: the test takes longer to write than the feature and breaks on the next redesign. TesterArmy removes the scripting layer entirely rather than generating scripts for you to maintain, so a moved button does not fail the build. Compared with other agent-based testers, the differences are the blocked verdict, the full transcript and browser logs on every run, first-class mobile including Expo, and the MCP server, which most tools in the category do not offer. There is a free plan with no card, so the honest answer is to run your own checkout flow on both and compare the transcripts.
Product and engineering teams at software companies from seed to Series C that ship web or mobile apps often and do not have a QA team, or have one that cannot keep up. The people who set it up are product engineers, founders and CTOs. Teams building with React Native or Expo are a strong fit because mobile end-to-end testing is where scripted tools hurt most. QA engineers use it to stop maintaining brittle scripts and spend the time on test strategy instead.
Szymon Rybczak and Oskar Kwaśniewski are React Native core contributors who worked together across four companies, shipping web and mobile products used by millions. In every one of them QA was the slowest part of shipping, and end-to-end tests were the first thing dropped under deadline pressure. They built an agent that tests like a user because that was the only version of testing they had seen teams keep using. TesterArmy is part of Y Combinator’s Spring 2026 batch.
TypeScript throughout. The agent drives real browsers through Playwright and runs on frontier vision and language models, selected per task. The CLI is built on the Vercel AI SDK and Playwright, and the MCP server speaks Streamable HTTP with OAuth sign-in. The web app is Next.js on Vercel. The open-source parts, the Scout API-testing CLI and the unbox-ai trace reader, are MIT licensed on GitHub.
We have collected here some useful links to help you find out if TesterArmy is good.
Check the traffic stats of TesterArmy on SimilarWeb. The key metrics to look for are: monthly visits, average visit duration, pages per visit, and traffic by country. Moreoever, check the traffic sources. For example "Direct" traffic is a good sign.
Check the "Domain Rating" of TesterArmy on Ahrefs. The domain rating is a measure of the strength of a website's backlink profile on a scale from 0 to 100. It shows the strength of TesterArmy's backlink profile compared to the other websites. In most cases a domain rating of 60+ is considered good and 70+ is considered very good.
Check the "Domain Authority" of TesterArmy on MOZ. A website's domain authority (DA) is a search engine ranking score that predicts how well a website will rank on search engine result pages (SERPs). It is based on a 100-point logarithmic scale, with higher scores corresponding to a greater likelihood of ranking. This is another useful metric to check if a website is good.
The latest comments about TesterArmy on Reddit. This can help you find out how popualr the product is and what people think about it.
Do you know an article comparing TesterArmy to other products?
Suggest a link to a post with product alternatives.
Is TesterArmy good? This is an informative page that will help you find out. Moreover, you can review and discuss TesterArmy here. The primary details have been verified within the last quarter. So they could be considered up to date. If you think we are missing something, please use the means on this page to comment or suggest changes. All reviews and comments are highly encouranged and appreciated as they help everyone in the community to make an informed choice. Please always be kind and objective when evaluating a product and sharing your opinion.
