#046 • 11 November 2025

Contents of this Newsletter

🤷🏼‍♂️

The What, Why and How QA Mode works

⏰ : 5 min

✅

Quash Tests Your Apps

⏰ : 4 min

🔔

What We Shipped Last Week

⏰ : 2 min

📓

Blog Picks

⏰ : 10 min

🗞️

Weekly AI News

⏰ : 7 min

MAIN STORY

Why We Built QA Mode?

If you have worked on a mobile app recently, you already know how unpredictable it has become. The number of devices and OS combinations keeps growing and that is only part of the problem. As experiences become increasingly personalized and teams run continuous experiments, what one user sees can be entirely different from another user (and both experiences can be correct). Today, mobile development moves at a pace that those systems were never designed to handle. Apps are updated weekly, tested across countless devices, and expected to perform perfectly under unpredictable conditions. The margin for error has disappeared.

The data reflects what teams feel every day. The average mobile app loses 95% of new users within 90 days, and most retain less than 10% after the first month. A crash, lag, or layout issue directly erodes trust and retention. In mobile, quality failures are business failures.

Over the years, testing has evolved from manual test cases to intelligent agents that can interact with apps directly. Mahoraga, the engine that powers Quash, is one such innovation. It can open a mobile app, scroll through screens and validate flows using natural language prompts, all driven by intent. This approach works brilliantly for verifying that core features behave as expected.

But intent alone can only take testing so far. Real QA is about uncovering where and how something breaks. That requires a mindset built around failure, not functionality. And that gap is exactly what led us to create QA Mode.

The Thinking Behind QA Mode

Mahoraga, by design, is a liberal system. It interprets human intent, executes natural language prompts, and behaves as a user would: navigating screens, filling forms and verifying that the app’s primary functions hold up. This makes it perfect for functional and regression testing, the phase that confirms whether a feature works at a broad level.

However, traditional QA goes further. Testing is about exploring what users might do wrong or where they get lost. Intent-driven testing can confirm 80-90% of the functional paths, but the remaining 10% (the edge cases, unexpected inputs, and abnormal flows) are where real quality is proven. QA Mode was built to enable that deeper layer. It introduces a stricter, failure-driven layer within Mahoraga, one that demands prompt precision, specificity in steps and clearly defined outcomes. In this mode, every instruction must be descriptive enough for the system to replicate edge conditions and test how the app behaves when something unexpected happens.

A study, “AI-Driven Tools in Modern Software Quality Assurance,” (June 2025) found that AI systems with stricter, well-defined execution had only 8.3% flaky runs, showing that precision makes automation more reliable.

How QA Mode Works (and Why It Matters)

QA Mode is built around the same three stages that define every QA workflow: Planning, Execution, and Reporting. What changes in this mode is how each stage is approached.

In Planning, testers define structured test steps that describe both intent and expectation. Instead of simply saying “tap login,” a QA Mode prompt might specify, “tap login after entering an invalid password, and verify that an error message appears.” This level of clarity allows the system to test not just the happy path but also the potential points of failure. During Execution, Mahoraga operates in a stricter state. Every action is carried out as described, with no interpretation or deviation. The engine follows explicit commands, ensuring that tests cover every defined edge condition. It obeys the test structure fully, mirroring the precision required in traditional QA. Finally, in Reporting, every run generates a detailed context trail combined with AI-powered insights that analyze what went wrong and why. This reporting framework turns each failure into a data point that can be studied and reused, helping teams understand not just what broke, but the conditions that led to it.

The outcome is a mode that complements the existing power of Mahoraga rather than replacing it. The intent-driven engine handles the broad validation of functionality, while QA Mode pushes the same engine into a disciplined search for edge-case behavior. Together, they cover both sides of the QA process confirming what works and uncovering what doesn’t.

The future of QA will be defined by how deeply tests understand failure. Mahoraga made it possible for a system to use mobile apps like a human. QA Mode makes it possible for that system to question those interactions with the same curiosity and discipline as a tester.

Your Apps, We Test

Each week, we put our execution engine to work on real apps.

This week, we put our execution engine to work on Eloelo, the live social entertainment app that brings creators and audiences together through interactive video and audio livestreams. Using a simple natural language prompt, we asked Quash to open the app, head to the Zones section, and navigate through different live zones. It joined the first zone from the list, returned, and then joined another one. Once done, Quash went back to the zone list to confirm that both joined zones were now displayed correctly, ensuring that the app’s join and navigation flow worked seamlessly from end to end.

Want us to test your app next? Drop us a mail at hello@quashbugs.com

What Did We Ship?

QA Mode

v1.0.5 | November 11, 2025

This week, we shipped QA Mode, a new way to bring structure and discipline to your test runs. It lets you define every step of a flow in detail, so the agent follows your exact instructions, one by one.

While Mahoraga is built to understand intent and perform actions naturally, QA Mode shifts its focus from just making things work, to actively testing where they might fail. It enables stricter, more detailed prompts that mimic how traditional QA engineers approach validation—checking flows, edge cases, and breakpoints with precision.

With QA Mode, Mahoraga transitions from an intent-driven assistant to a disciplined tester. It’s the difference between proving a flow works and ensuring it can’t break.

Check out our Recent Blogs

News

This Week In AI

China Bans Foreign AI Chips from State-Funded Data-Centres

China has issued guidance requiring that new data-centre projects receiving any state funding must use only domestically-manufactured AI chips. Projects less than 30% complete are being told to remove all foreign chips or cancel orders, while more advanced sites will be assessed case-by-case. This directive adds to existing efforts to reduce reliance on U.S.-made AI hardware and boost China’s domestic chip sector.

OpenAI Offers One-Year Free ChatGPT Go Access in India

OpenAI has announced that all users in India will receive one year of free access to its ChatGPT Go plan starting November 4, 2025. The plan includes faster response times, early access to new features, and improved mobile performance. This move marks OpenAI’s first large-scale regional rollout of a premium plan at no cost, aimed at expanding adoption and usage in one of its fastest-growing markets.

Snapchat Partners with Perplexity AI for In-App Search

Snapchat has partnered with Perplexity AI to bring an AI-powered search feature directly into the app, enabling users to ask questions and receive sourced, conversational answers. The deal, valued at around $400 million for one year, will make Perplexity’s answer engine a core part of Snapchat’s global rollout beginning in early 2026.

All Newsletter Issues | Security | Contact Us | Blogs

How Our QA Mode Changes Everything in Mobile Testing 😎

Product

Solutions

By role

By use-case

By industry

Resources

Comparison

Company

Product

Solutions

By role

By use-case

By industry

Resources

Comparison

Company

Get Quash

Product

Solutions

By role

By use-case

Resources

Comparison

Company

Get Quash