VeriWasp is an AI powered playtester and eventually full testing suite for your product with manuall runs or added into your existing CI/CD. You paste a URL, pick a template (signup, checkout, invite-teammate, that kind of thing) or just describe your own flow in plain English, and it drives a real headless browser through it, clicking, typing, asserting, the same way a person would. Every run gets a full accessibility scan and produces a public report to share. No login needed to view it.
Big quality-of-life upgrade for everyone hunting down those sneaky bugs!We just shipped the Full Trace Viewer: a deep-dive debugging panel for every run.
What it gives you:
Complete network tab per step: every request/response with headers, sortable table, URL filtering, and full request/response bodies (up to 5MB)
Side-by-side DOM snapshots: actual reconstructed DOM + computed styles before and after each step (not just screenshots)
Works on public/shared reports too: perfect for collaborating with teammates or customers
Hi all! Creating another designated place to share new major features and updates as we lead up to launch! Feel free to use this space to request features as well!
Hi everyone! I am starting this thread as a designated place to discuss the features and offerings of VeriWasp. If you give it a try, let me know what's broken, missing, or what's working well. I want to ensure this product is useful and robust so don't hold back!
Otari by Mozilla AIControl Your LLM Infrastructure
Promoted
Maker
📌
Hey everyone! I built VeriWasp after too many "it works on my machine" but then prod breaking after a deploy at midnight and not finding out until a support email the next morning. No QA team, no time to click through every flow by hand before every deploy.
You paste a URL, either pick a template (signup, checkout, invite-teammate, that kind of thing) or just describe your own flow in plain English, and it drives a real headless browser through it, clicking, typing, asserting, the same way a person would. Element matching is AI-primary (Claude Haiku, live, on every step) with deterministic text/attribute matching as a fallback, so a redesign doesn't quietly break your test suite the way brittle CSS selectors do. It also offers AI chat about a run to fix things faster with the run info in context.
Every run also gets a full accessibility scan (axe-core, real WCAG checks) for free, and produces a public report: video replay, a screenshot per step, every console error tied to the step that caused it, and per-step timing. No login needed to view it. It's meant to be the thing you paste into a launch thread as proof something works, not just a claim.
It also has a trigger URL per flow, so it plugs into your CI/CD (GitHub Actions, a Fly release command, whatever) if you want "does checkout still work" checked on every deploy.
Pricing is credit packs, not a subscription and starts at 50¢/run, 10 free credits on signup, no card needed to try it. I didn't want billing anxiety to be the reason someone doesn't run one test.
Hey everyone! I built VeriWasp after too many "it works on my machine" but then prod breaking after a deploy at midnight and not finding out until a support email the next morning. No QA team, no time to click through every flow by hand before every deploy.
You paste a URL, either pick a template (signup, checkout, invite-teammate, that kind of thing) or just describe your own flow in plain English, and it drives a real headless browser through it, clicking, typing, asserting, the same way a person would. Element matching is AI-primary (Claude Haiku, live, on every step) with deterministic text/attribute matching as a fallback, so a redesign doesn't quietly break your test suite the way brittle CSS selectors do. It also offers AI chat about a run to fix things faster with the run info in context.
Every run also gets a full accessibility scan (axe-core, real WCAG checks) for free, and produces a public report: video replay, a screenshot per step, every console error tied to the step that caused it, and per-step timing. No login needed to view it. It's meant to be the thing you paste into a launch thread as proof something works, not just a claim.
It also has a trigger URL per flow, so it plugs into your CI/CD (GitHub Actions, a Fly release command, whatever) if you want "does checkout still work" checked on every deploy.
Pricing is credit packs, not a subscription and starts at 50¢/run, 10 free credits on signup, no card needed to try it. I didn't want billing anxiety to be the reason someone doesn't run one test.