llms.txt
Compare with other sites' llms.txt files →# ClankerView
> AI user testing. ClankerView sends AI agents, each with its own persona, to use a website in a real browser while thinking out loud. Each agent writes a review of what it experienced, and you keep the recording of its visit.
ClankerView is for indie hackers, founders and product managers who want to see how people might experience their site before real users do. Paste a URL and agents scroll, click, fill in forms and sign in the way a human tester would. They react to the actual interface, not to a description of it, which is the difference from asking a chatbot about a page's HTML.
- Any website with a URL works: SaaS products, landing pages, onboarding flows, shops, staging sites and localhost tunnels.
- Agents can test behind a login, with credentials in their instructions or with session cookies pasted in. Cookies are deleted when the run finishes.
- Personas are cast to suit the site, or described by you ("a skeptical enterprise IT manager"), and each agent can get its own instructions and screen size, from phone to desktop.
- You can watch each agent live. Afterwards every agent's review stays with its recording, split into chapters, and insights list the issues several agents ran into.
- Coding agents like Claude Code, Cursor and VS Code can start reviews and read them through an MCP server, and scripts and CI through an API, at the same prices.
- Most reviews take 2 to 6 minutes. Each agent's review can be shared with a link, and reviews nobody opens for a year are deleted, after a warning.
Pricing: new accounts get 5 credits free, no card needed, and 1 credit is $1. Each agent's review costs 2 credits, or 3 credits for an extended review that browses twice as long. It's charged when the review is written, and an agent that fails isn't charged. Extended reviews and several agents at once open up with the first top-up. Credits come in packs (10 credits for $10, 20 credits for $19, 50 credits for $45, 100 credits for $80) and don't expire.
## Pages
- [ClankerView](https://clankerview.com/): what it does, real recorded reviews, pricing and frequently asked questions
- [Help center](https://clankerview.com/help): guides for using ClankerView
- [Blog](https://clankerview.com/blog): articles on user testing, usability and AI agents
## Help
- [Signing in with passkeys](https://clankerview.com/help/signing-in-with-passkeys.md): Sign in with Touch ID, Face ID, Windows Hello or your password manager instead of an emailed code.
- [Testing behind authentication](https://clankerview.com/help/testing-behind-authentication.md): Let agents review the pages behind your sign-in, with a test account's password, its cookies or its local storage.
- [The API and MCP server](https://clankerview.com/help/api.md): Start reviews and read what the agents found from your own scripts, CI and coding agents.
## Blog
- [How many users do you need for a usability test?](https://clankerview.com/blog/how-many-users-usability-test.md): Five, mostly. Where the famous number comes from, when it’s wrong, and how to spend a small testing budget well.
- [How to user test your app without recruiting testers](https://clankerview.com/blog/user-testing-without-recruiting-testers.md): You don’t need a panel, a budget or a research team to watch someone struggle with your product. Seven ways to find usability problems this week.
- [Can AI really use your app like a real user?](https://clankerview.com/blog/can-ai-use-your-app-like-a-real-user.md): Mostly, for the first few minutes of a stranger’s visit. What AI agents get right when they click through a website, where they fall down, and what that means for how you use them.
- [Send users* to review every deploy automatically](https://clankerview.com/blog/review-every-deploy-automatically.md): *AI ones. How to have an agent click through every preview deploy and tell you what got harder to use, from CI or straight from Claude Code.
## Legal
- [Terms of service](https://clankerview.com/terms)
- [Privacy policy](https://clankerview.com/privacy)
## Optional
- [The help center in one file](https://clankerview.com/llms-full.txt)