Nanosaur2-670M
well9472 · Kit
Makes illustrations from text prompts with a small 670M-parameter diffusion model, run through ComfyUI. Built as a low-budget research model for people who wan…
Runs end-to-end UI tests for Android, iOS and web apps from plain YAML files, driving the app from outside. For developers and QA teams who want quick test runs on a laptop and in CI.
Speedrun is an open-source project on GitHub in the Developer category, listed by Mindket. It is free. This is version 1, last updated on . Install it from GitHub as its README says. Get Mindket’s setup notes on your Mindket account, as a zip.
brew install hubastard/tap/speedrun
Speedrun is a test runner for mobile apps and websites. You write each test as a short YAML file, and it taps, types, scrolls and checks the screen from outside your app. Nothing is added to the app itself, so it works on any build.
Mobile app developers and QA engineers who want UI tests that start quickly and fail less often. Teams that already write Maestro flows can reuse them.
| Industry | Who uses it | How they use it | Example |
|---|---|---|---|
| Mobile app development | App developers | Write flows that tap through sign-in and other screens on Android and iOS | A developer checks after each change that the Sign in flow still ends on the Welcome screen |
| Quality assurance | QA engineers | Keep a suite of YAML flows and run it with JUnit or HTML reports | A tester runs a whole folder of flows and reads the HTML report |
| Web product teams | Front-end developers | Run the same kind of flow in Chrome or Firefox against a page | A developer runs a login flow in headless mode before a release |
| Test automation | Teams on Maestro | Switch the command name and keep existing flow files | A team replaces maestro test with speedrun test on its flows folder |
Teams that only write unit tests, or that test desktop apps, can skip it.
| Use case | Who | What they get back |
|---|---|---|
| Check a sign-in flow | App developers | A result for each step, plus what was on screen and saved logs when one fails |
| Run a suite in CI | QA and release teams | JUnit and HTML reports, with flows sharded across devices |
| Move off Maestro | Teams with existing flows | The same flow files, config.yaml, tags and reports after a one-word change |
You need a way to install it and a device to test on.
brew install hubastard/tap/speedrun
Write a flow file, then run it. Each step appears in the terminal as it runs, and a failed step shows what was on screen and where the screenshot and logs went.
appId: com.example.app --- - launchApp - tapOn: "Sign in" - inputText: "user@example.com" - tapOn: "Continue" - assertVisible: "Welcome"
speedrun test login.yaml -e PASSWORD=secret speedrun test -p ios login.yaml -e PASSWORD=secret speedrun test flows/ --format junit --output report.xml
| Tool | What it does |
|---|---|
| launchApp | Opens the app, and starts clean when clearState is set to true |
| tapOn | Taps the element that shows the given text |
| inputText | Types the given text into the field |
| assertVisible | Checks that the given text is on screen |
| speedrun test | Runs one flow file or a whole folder, on Android by default or on iOS with the -p ios option |
| speedrun fast-boot | Boots an emulator or simulator with the apps and services your tests never need switched off |
A web flow points at its page through a url key, in the place where an app flow has appId. Run it with the headless option for Chrome, or add the Firefox device option.
You write the five-step flow above. It opens the app, taps Sign in, types an email address, taps Continue and checks that Welcome is on screen. You run it with speedrun test and your file name. If the check fails, the terminal shows what was on screen and where the screenshot is.
Speedrun uses your app the way a person would. It taps buttons, types text and reads the screen, then tells you which step broke. A first small thing to try is the five-step flow above against an app on an emulator.
Underneath, it works through the platform's accessibility layer, so release builds work and no SDK goes into your app. Each step starts when the screen stops changing instead of after a fixed sleep, and screens that never settle, such as spinners, do not hold it up. Its Android and iOS agents ship inside the one file, and it controls Chrome and Firefox directly, so there is no JVM or WebDriver to install. The trade-off is that Maestro's AI commands are not supported.
Runs end-to-end UI tests for Android, iOS and web apps from plain YAML files, driving the app from outside. For developers and QA teams who want quick test runs on a laptop and in CI.
Made by hubastard on GitHub. These notes are Mindket's, written from its README; the project's own files and license are in its repository.
What it is: Speedrun is a test runner for mobile apps and websites. You write each test as a short YAML file, and it taps, types, scrolls and checks the screen from outside your app. Nothing is added to the app itself, so it works on any build.
Who it suits: Mobile app developers and QA engineers who want UI tests that start quickly and fail less often. Teams that already write Maestro flows can reuse them.
Industry | Who uses it | How they use it | Example
Mobile app development | App developers | Write flows that tap through sign-in and other screens on Android and iOS | A developer checks after each change that the Sign in flow still ends on the Welcome screen
Quality assurance | QA engineers | Keep a suite of YAML flows and run it with JUnit or HTML reports | A tester runs a whole folder of flows and reads the HTML report
Web product teams | Front-end developers | Run the same kind of flow in Chrome or Firefox against a page | A developer runs a login flow in headless mode before a release
Test automation | Teams on Maestro | Switch the command name and keep existing flow files | A team replaces maestro test with speedrun test on its flows folder
Teams that only write unit tests, or that test desktop apps, can skip it.
Where to use it:
Use case | Who | What they get back
Check a sign-in flow | App developers | A result for each step, plus what was on screen and saved logs when one fails
Run a suite in CI | QA and release teams | JUnit and HTML reports, with flows sharded across devices
Move off Maestro | Teams with existing flows | The same flow files, config.yaml, tags and reports after a one-word change
How to set it up: You need a way to install it and a device to test on.
brew install hubastard/tap/speedrun
How to use it: Write a flow file, then run it. Each step appears in the terminal as it runs, and a failed step shows what was on screen and where the screenshot and logs went.
appId: com.example.app
---
- launchApp
- tapOn: "Sign in"
- inputText: "user@example.com"
- tapOn: "Continue"
- assertVisible: "Welcome"
speedrun test login.yaml -e PASSWORD=secret
speedrun test -p ios login.yaml -e PASSWORD=secret
speedrun test flows/ --format junit --output report.xml
Tool | What it does
launchApp | Opens the app, and starts clean when clearState is set to true
tapOn | Taps the element that shows the given text
inputText | Types the given text into the field
assertVisible | Checks that the given text is on screen
speedrun test | Runs one flow file or a whole folder, on Android by default or on iOS with the -p ios option
speedrun fast-boot | Boots an emulator or simulator with the apps and services your tests never need switched off
A web flow points at its page through a url key, in the place where an app flow has appId. Run it with the headless option for Chrome, or add the Firefox device option.
Example: You write the five-step flow above. It opens the app, taps Sign in, types an email address, taps Continue and checks that Welcome is on screen. You run it with speedrun test and your file name. If the check fails, the terminal shows what was on screen and where the screenshot is.
For newcomers and experts: Speedrun uses your app the way a person would. It taps buttons, types text and reads the screen, then tells you which step broke. A first small thing to try is the five-step flow above against an app on an emulator.
Underneath, it works through the platform's accessibility layer, so release builds work and no SDK goes into your app. Each step starts when the screen stops changing instead of after a fixed sleep, and screens that never settle, such as spinners, do not hold it up. Its Android and iOS agents ship inside the one file, and it controls Chrome and Firefox directly, so there is no JVM or WebDriver to install. The trade-off is that Maestro's AI commands are not supported.
Write the test as a YAML flow file
Speedrun drives the app from outside
Each step waits for the screen to settle
A failure saves a screenshot and logs
Reports come out as JUnit or HTML
Good to know:
Title: Time for the same user journey on Android
Speedrun: 20 s
Appium: 59 s
maestro-runner: 66 s
Maestro: 127 s
git clone https://github.com/hubastard/speedrun.git
Then follow the setup in its README. Nothing in this kit runs by itself.
README.md 6 KB
Listed GitHub projects in the same category that share the most topics with this one.
| Product | Stars | Activity | Language | License | Last change |
|---|---|---|---|---|---|
| Speedrun This page | 100 | Active | Odin | Apache-2.0 | |
| Tester Army e2eShares e2e-testing · end-to-end-testing | 13,443 | Active | TypeScript | Apache-2.0 | |
| agent-desktopShares cli · mcp · ai-agents | 19,705 | Active | Rust | Apache-2.0 | |
| kanzoShares cli · mcp · ai-agents | 1,045 | Active | Rust | AGPL-3.0 | |
| qwen-codeShares cli · mcp | 29,845 | Active | TypeScript | Apache-2.0 | |
| ECC by affaan-mShares mcp · ai-agents | 299,210 | Active | JavaScript | MIT |
well9472 · Kit
Makes illustrations from text prompts with a small 670M-parameter diffusion model, run through ComfyUI. Built as a low-budget research model for people who wan…
Waqf-AI · Kit
Reads handwritten Arabic text from images and returns it as text. A fine-tuned PaddleOCR-VL model for people digitizing Arabic handwriting.
Wan-AI · Kit
Makes video clips from a text prompt or an image plus text, at up to 720P and 24 frames per second, in the Diffusers format for a single consumer GPU.
An MCP server and CLI by morluto that lets your coding agent study binaries, desktop and Android apps, websites and firmware on your own computer, and back eac…
vercel-labs · Kit
Enables AI agents to read and interact with desktop applications on macOS, Windows, and Linux using accessibility trees instead of screenshots. It provides a C…
emanuele-em · Kit
Runs a kanban board in your browser and an MCP server from one small program, so you and your coding agents share one task list. It needs no account and no ser…