Record manual QA flows, get E2E test code that fits your repo
TLDR: Desktop app for E2E web test generation, built at JetBrains (closed beta). Record the flow in a built-in browser - the agent matches it with your existing codebase, then writes a test that passes, not a draft to debug. Devs use AI to ship more code. That code still needs testing. If your team writes E2E tests by hand, you have a problem - same QA capacity, way more surface to cover. AI agents can write E2E test code, but you're stuck describing flows in text - the agent clicks around via Playwright MCP, takes wrong turns, you re-prompt, retry. 30 minutes for a flow you could click…
In plain words
Qure is a desktop application for generating end-to-end web test code by recording user flows through a built-in browser. Built by JetBrains and currently in closed beta, it lets QA teams and developers record actual interactions with their product, then automatically generates test code that integrates with their existing codebase. Unlike text-based AI test generation that requires prompt engineering and often produces drafts needing debugging, Qure converts recordings directly into working tests. It's designed for teams writing E2E tests manually who need to increase test coverage without proportionally expanding QA capacity.
written from the facts on this page · September 2026
From the sources
In the maker’s words, at launch
TLDR: Desktop app for E2E web test generation, built at JetBrains (closed beta). Record the flow in a built-in browser - the agent matches it with your existing codebase, then writes a test that passes, not a draft to debug. Devs use AI to ship more code. That code still needs testing. If your team writes E2E tests by hand, you have a problem - same QA capacity, way more surface to cover. AI agents can write E2E test code, but you're stuck describing flows in text - the agent clicks around via Playwright MCP, takes wrong turns, you re-prompt, retry. 30 minutes for a flow you could click through in 30 seconds. Qure works differently. You record the scenario in Qure's built-in browser by just using your product. The AI turns that recording into code. No prompt engineering, no MCP setup, no explaining your repo in chat - point it at your project and go. Beyond recording, you can also refactor tests, update them, or write new ones from a description. What keeps the AI output grounded: - We match the recording against your codebase - find your page objects, helpers, constants and feed them to the agent instead of hoping it figures out your repo - When agent runs the test, it reads real failure output, fixes with actual error and app context This is a closed beta of an experimental product. Web only, works best with Playwright. If your project has a few dozen tests - Claude Code will honestly get you there. Qure makes a difference on larger codebases with existing test infrastructure. 5-min demo: https://www.youtube.com/watch?v=4CZw4bSSDCE Try the beta: https://quretests.com Happy to answer any questions about the approach, product, or where it breaks - I'm the dev on the Qure team. Egor (@250xp), who leads the project, is in the thread too.
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 26d ago · cactuscompute.com
Launched alongside, March 2026
the whole month →
Switch from ChatGPT to Claude with import memory feature
AI · Mar 2026 · claude.com
