Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Android CLI: Build Android apps 3x faster using any agent (googleblog.com)
272 points by ingve 19 hours ago | hide | past | favorite | 113 comments
 help



The install command shown for Windows is 404.

`curl -fsSL https://dl.google.com/android/cli/latest/windows_x86_64/inst... | bash`

The URL shown for individual OSs work, but the script errors for me.

`curl.exe -fsSL https://dl.google.com/android/cli/latest/windows_x86_64/inst... -o "%TEMP%\i.cmd" && "%TEMP%\i.cmd"`

I manually downloaded the exe, but it say socket error. vibe coding is going strong!


Goggles Android tooling has been like this forever, nothing to do with AI.

Ah, ok, no worries then, here I thought they didn't care about engineering quality or tooling that works just recently, but turns out they never did! Thanks :)

I got a workaround a la GH Copilot:

<pre>

> android skills list

Picked up JAVA_TOOL_OPTIONS: -Djava.net.useSystemProxies=true

</pre>


I honestly have no idea what is going on. Lots of broken things in what's supposed to be front products for Google and other "high name" brands. I don't get it: Where is everybody? Is there no one there? Are these companies really dead inside?

Same for Microsoft. Redirects to the void, 5-level-deep sign-in prompts, "contact your administrator" who doesn't exist...

Maybe it's a size thing.


>Google collects usage data for the Android CLI, such as commands, sub-commands, and flags used. This data does not include custom parameters or identifiable information. This information helps improve the tool and is collected in accordance with Google's Privacy Policy.

>https://policies.google.com/privacy

>Disable Android CLI metrics collection by using the --no-metrics flag.

No thanks, is there no env variable for this? Doesn't Google have enough data already?


Android CLI can write a tool that wraps android-cli and automatically passes the flag based on an env variable.

How would Google have enough data about a brand new product without collecting that data?


> How would Google have enough data about a brand new product without collecting that data?

They wouldn't. But on the other hand, they probably have a large amount of in-house Android app developers on whom they can conduct such metrics collection. I wouldn't expect outsiders to have vastly different workflows, because when you get out of the happy path with Android all you get is pain.


Outsiders probably do have vastly different workflows. Google internally love to stick Bazel on everything and that's quite different (and overly complicated) compared to the usual Gradle route.

`alias android-cli='android-cli --no-metrics'`

Uh do aliases load in non-interactive shells?

Create a wrapping binary instead

    mkdir -p ~/.local/bin
    printf '#!/usr/bin/env sh\nexec android-cli --no-metrics "$@"' > ~/.local/bin/android-cli
    echo 'PATH="$HOME/.local/bin:$PATH"' >> ~/.zshenv

I'm pretty sure this will just call itself in a loop. You need to use the absolute path to the wrapped binary to distinguish it from the wrapper.

Aah! Yes absolutely right! Using `exec command android-cli` would work I believe

Also it's not a binary :-)

You could export BASH_ENV to have Bash processes source a given file at startup.

Zsh has .zshenv, and Fish just has config.fish for everything with the ability to guard certain things within it to login only or non-interactive only.


I wish the same thing existed for Apple.

Everything I do for macOS/iOS is already without Xcode but it's a pain in the ass to keep up with changes, and there are things I haven't figured out yet (like AUv3).


> Everything I do for macOS/iOS is already without Xcode

Doesn't Xcode allow you to plug in agents like VS Code does?


I don't know, I don't use/want to use Xcode.

What I'm interested is in the CLI tool for my own use, not necessarily for agents.


Come on now, you're going to need Xcode at some point during development.

No, you really don't. Apple would like you to use it (god knows why), but it's not necessary.

It does.

Wow. Thanks for this update. It streamlined a lot of tasks.

Apart from this, next step will be to add suport for building android apps on the android phones itself. No desktop needed.Building on the laptop with agents and installing the build in the phone and testing doea not seem AI native. If everything can run on my android phone, development cycle will speed up.


you already could! just install Termux, npm install your favourite agent harness (pi for one has explicit Termux support, but its AGENTS.md works just fine with Claude Code for example - https://github.com/badlogic/pi-mono/blob/main/packages/codin...), and say you want an android app. It problem solves for a bit, then spits out an apk out to your Downloads folder.

Let me try this. Last year this was a dream. Can't belive we are so close to automate all of this.

My major issue last time was providing the feedback to the agent by running the apk on phone i.e, pass the debug log from the apk back to agent so it can iterate on it without me providing any input.


ask the agent to run adb log so it can read it for itself

Also coding agents will happily compile android applications (of maximum complexity) via Github Actions where you can just pick them up with Obtainium. No PC needed

What is obtainium.

An android app that tracks releases to install the latest versions of apps directly from github.

Also it works with private repos too if you provide a personal access token (fine-grained) in the Obtainium app settings. Just make sure to "release" (on the Releases tab) the .apk file on the GitHub repo and tag it latest.

I use 'just' (command runner) and the 'gh' CLI to automate this:

    # Build and publish a GitHub release
    release: apk
        VERSION="v$(date +'%Y.%m.%d')-$(git rev-parse --short HEAD)"; \
        APK="build/app/outputs/flutter-apk/app-release.apk"; \
        REPO="$(git config --get remote.origin.url | sed -E 's#.*github.com[:/](.*)\.git#\1#')"; \
        echo "Releasing $VERSION to $REPO"; \
        git tag "$VERSION" 2>/dev/null || true; \
        git push origin "$VERSION"; \
        gh release create "$VERSION" "$APK" \
            --repo "$REPO" \
            --title "$VERSION" \
            --notes "Automated release for $VERSION" \
        || gh release upload "$VERSION" "$APK" --repo "$REPO" --clobber

You actually can right now on termux.

This is a good step forward, but keep in mind the claimed gains are about "project and environment setup", not the tasks you deal with on a daily basis in an existing project.

Taking screenshots, optionally with component borders highlighted, and operating the UI with element names like "button1" instead of tap 200,30 looks useful. If I could get it to work.

Android Studio on its deathbed. Just release VSCode plugin and kill it for good, it has been a buggy, slow mess for the last 3 years or so.

I can't be the only one with a severe aversion to language/platform specific IDEs. "Oh no, you can't write Python in Vim, you need to get Pycharm!" "iOS apps? Don't try anything except XCode."

IDEs are (were? :-/) a very personal choice. Maybe I'm an AI but I would have loved a CLI-centric workflow. It would have kept options open.


If you think Android Studio has been a bug in the last 3 years, you haven't used it back in the days of Android Studio Beta. It has come a long way.

Must be a web dev who adamantly uses VSCode for everything

Do you still use IDEs?

I haven't used an IDE since December.


Debugging and local testing is the main remaining use case of IDEs. Especially so for mobile apps where you need to manage one or more emulators.

Not on Android, where debugger works once in a blue moon.

`android docs` is the superpower we need for everything. NPM / pnpm should have similar `npm docs` that would allow humans and agents to search for type-signatures and JSDocs.

It is so annoying that each agent has its own ideas where it tries to get the docs, usually by blindly grepping.


> 3x faster

Because the real bottleneck is really the velocity of development, right next to keeping the codebase small - right guys?


Code-generation velocity, and the absolute number of lines churned are 2 metrics AI tools can irrefutably benchmaxx, so the vendors highlight them.

Oh absolutely - the old saying about only having a hammer and seeing each problem as a nail applies.

On the positive side, they've decided to come down from "superintelligence", "superchanging workflows" and other bullshit to the actual feature - 3x speed of text generation. Which is not quite the problem that needed to be solved in software engineering, but as you said yourself...


CLI in a product name now means LLM agent TUI specifically and not just, as I would have expected, any kind of Command Line Interface? And usually there is barely a CLI included at all and you are expected to mostly launch the full TUI with its own embedded readline-loop rather that use the CLI?

This is a CLI, the example shows the Gemini TUI using it

Efficiency claim: 70% less token usage, 3x faster task completion in internal testing. Even if that's marketing-inflated, the structured CLI commands replace a lot of trial-and-error. Gonna try it.

Let's see if even mid/big companies with tons of resources, with AI and the right tooling will continue to write webview-apps or, even worse, use some kind of multi target wrapper.

> Your agents perform best when they have a lightweight, programmatic interface to interact with the Android SDK and development environment.

F you google. Me too. Why didn't we get a sane way to build android apps before you had to please chatbots?


Damned if you do. Damned if you dont.

Google has been neglecting Android for years with subpar tooling and ridiculous development practices.

If you think Android tooling is subpar wait until you try iOS.

Apple deliberately makes them shitty, big difference.

Damned if you don't, damned if you do fifteen years later for an entirely different reason.

This is great. We also need a tool to expose source jars to agents so they don’t need to compress. There’s a lot of Compose overloads that Claude just guesses at. I built something internally but it needs polish and Claude really struggled with the deep Gradle integration.

great! now let me know when your official app store transparently alerts users when an app they were using was sold to a third-party adtech surveillance company, please :)

How can I use this official android skill with Claude code?

Is there any step by step process or guidance on it?


Looks like the docs start here: https://developer.android.com/tools/agents/android-skills#us...

There's a link to a repo or you can use the CLI tool to init the skills


But can I publish an app without having to share my ID? I want an ecosystem that doesn't require it.

It's not just your ID; it's your address, phone number, and the list goes on.

Absolutely not. That would be crazy.

Zapstore or Obtanium...

Now please let us install the apps just as easily

downloading an APK and opening it is already about as easy as it gets. the only thing easier would be for someone else to do it for you

You're forgetting the installation ("sideloading", what everyone else calls installation) restrictions they are about to deploy. It will be a significant hassle to install anything without Google's approval. Many F-droid apps are showing warning notices about this upcoming change.

Good, it shouldn't be two clicks for elderly people to install trojans on their phone that then drain their bank account. There should be some explicit confirmation that the user knows what they are doing and they are not being scammed. It is long overdue.

> Good, it shouldn't be two clicks for elderly people to install trojans on their phone that then drain their bank account.

And what makes you think that most scams involve fancy zero days/CVEs/hijacking the OS, and not simple social engineering?

You do not require a malicious apk to receive 2FA codes, or for the gullible user to read them aloud to the scammer. All phones come with an SMS and phone app.

You do not require a malicious apk to send transactions in banking apps (eg tricking people selling their product to send the money.)

You do not require a malicious apk to engage in a pig butchering scam, or to buy gift cards.

> There should be some explicit confirmation that the user knows what they are doing and they are not being scammed. It is long overdue.

I agree. Social engineering counters should have awareness raised by the governments. But blocking 3rd party apps for this is like using a cannon to shoot a mosquito. I'm not sure it makes the slightest of sense.


We can and should address more than one problem at a time.

Malicious APKs are a real problem that exists. I work tangentially in this space.

> But blocking 3rd party apps for this is like using a cannon to shoot a mosquito.

I’d agree, if that was what was going to happen. But it isn’t. Google is not going to block 3rd party apps.


> We can and should address more than one problem at a time.

Very much agree. Here in India, one of the big telecos has now rolled out a system where if you're on a call with an unknown number, OTPs are not sent to the phone till the call ends. IMO systems like this (or ironically - using OEM installed on device AI as a MITM to stop a call when an OTP is heard) are very good ideas.

> Malicious APKs are a real problem that exists. I work tangentially in this space.

Not doubting it for a moment. I've myself installed an app (that in my defense I pretty much suspected to be malware) that was malware. Even a few weeks ago I helped someone remove a hidden app that was draining their battery like anything (idk doing what, crypto mining or something I guess?). Ofc this app had accessibility permissions and would close settings if you tried to uninstall it.

On the flip side, I've also been stopped by my own phone to give accessibility permissions... to TapTap (a FOSS app by legendary developer quinny98) [1].

I should probably add - here in India, UPI scams use(d?) to be very common, let alone "giving someone your OTP" scams. I personally know someone very close who's lost a good bit of money, purely via someone social engineering them to hand over OTPs.

Even today, scamsters call and threaten a "digital arrest" (whatever the fuck that is) to unsuspecting victims. Presumably many hand over their money.

I have absolutely nothing against technical solutions. But IMO social education to never install apps from outside the play store, combined with "Digital Arrest does not exist" ads that the Indian govt is already running, are significantly stronger and resistant to much more things (like I mentioned - pig butchering or gift card scams).

I would be very curious if you had stats for how much is lost to scams via social engineering, vs malware. I asked Gemini (I can share the chat link via some private method of communication if you're interested), and apparently per IC3, it's 13.7B USD for social engineering, vs 1.57B USD for malware. If you have better data, I'd be happy to know more.

> I’d agree, if that was what was going to happen. But it isn’t. Google is not going to block 3rd party apps.

Perhaps I'm a cynical guy (which is true!), but I see zero reason to give google the benefit of doubt when it comes to control. I understand you're perhaps a googler (or you work on the same side) - nothing against it at all. Hardening is 100% helpful.

But companies famously like to increase revenue, and do not care about users. Every app on the play store (and btw there are a ton of scammy ones - I know because I get their ads on Youtube :) nets google some money. There's nothing stopping google from going "Actually we decided to stop all apk installs as people get scammed by them" tomorrow?

There is no fundamental reason to believe them beyond trusting them at their word. And there are many reasons to not believe them, unfortunately.

IMO, the old adage holds true - beating tech is hard, beating humans (with a wrench ;) is easy. Aka, XKCD 538.

1. https://github.com/KieronQuinn/TapTap 2. https://xkcd.com/538/


It is 1 click because the malware is on the play store already!

Both are problematic.

Think of the elders!

"This APK cannot be scanned and its safety cannot be verified. Learn more/go back" and "learn more" has a link that looks like nothing but is actually a button to actually install the app.

I can think of some easier things, for example popping up a dialog, pressing "install" and having my all actually be installed after that.


You're saying it should look like those damned browser certificate failure sites, with option to open the damn site hidden under button that looks like an unassuming link?

That's how it looks now.

I must say I'm quite disappointed.

I expected something useful for application development. All it offers is some wrapper around the basic Android setup command that LLMs are already good at. What, initial empty project creation now takes 5 minutes instead of 10? Big deal, who cares?

I had another hope awakening that at least skills might be useful. But except for a few migration recipes, there's nothing of value for day to day Android development.

Facit: I'll skip installing another Google app whose only purpose is more spying on me and keep developing Android apps the way I already do.

TLDR: Nothing to see here. Move on.


Can we have a web development CLI with web development skills?

Agents will allow human programmers to get what they've been begging for decades now: proper requirements and flexible, logical, tooling.

this has been my sort of big tent alignment with AI people. If I'm getting good CLI tooling that _actually works_ (or fixes to existing ones that have been busted forever) then I'm pretty happy.

Things that make systems more understandable to the LLMs ... usually make things more understandable for humans as well. Usually.

The biggest issue I've found is that vibed up tooling tends to be pretty bad at having the right kind of "sense" for what makes good CLI UX. So you still have awkward argument structures or naming. Better than nothing though


Its like major cities repairing their roads to incentivize autonomous vehicles to operate there. Win win for everyone.

Apart from pedestrians.

It never made sense to me why cars and pedestrians need to share the same spaces. Why can't we have more efficient walking routes that are away from cars?

Because cars took over the streets from pedestrians between 1900 and 1930 and no one noticed.

Hopefully when petrol hits $10 a gallon in the next few months more of the world will think about banning cars from high density areas.


Its already over $12 per gallon in Singapore. Let's see what happens.

if you have roads shared with pedestrians and cars (and bikes!) you can build denser cities.

I lived real downtown in Tokyo and my street was like "1.5" lanes wide (if cars were coming in both directions one basically needs to pull over and stop). I could just walk in the middle of the street. There was no sidewalk. No street parking of course. Cars would drive down at 15km/h or whatever, and slow to a crawl if people were in the street.

Straight lines are efficient walking routes, and ... well... that might involve just crossing the street directly! Every layer of grade separation gets in the way of that.

End result of all of this is less pavement to maintain, slower drivers (-> safer!), good walking and cycling conditions, etc etc etc.


Yes, we can do that by banning leisure cars trips from all dense areas.

What's that you say? Drivers are a major and rich political force and they will block such decisions?


Any textbooks or resources on getting better at naming things?

The Programmers Brain book was my go to



The Design of Everyday Things.

The conclusion I drew from that book is that I shouldn't be naming things.

I've been thinking the same thing lately. It's sorta frustrating that it required bots to force tech companies to make clean simple cli driven development workflows.

It's wild that it took AI to get half the companies on the planet to actually add reasonably priced APIs to their products so I don't have to puppeteer every damn thing with a flakey harness.

> Agents will allow human programmers to get what they've been begging for decades now: proper requirements and flexible, logical, tooling.

...and once this goal is finally reached the programmer will breathe a sigh of relief and then promptly be fired since now the machine can do the job as well as they could.


The tooling in 2026 is so easy you can do almost anything without AI very very quickly.

What tooling?

Frontend, backend, workflow engines, payments, idk all of it.

Like you can make a cross-platform react native app with expo in a day without any AI (a proper app, not a boilerplate). Same with many web apps.

Shadcn with themes and components, tailwind, temporal workflows, etc etc. The complexity of making apps was solved years ago.


We must live in different worlds. Even for professionals building high quality apps is hard. It's easier with AI, but it's still quite hard. And it was definitely harder without AI.

At the expense of no longer needing the human programmer...

[flagged]


What does this have to do with the Android CLI?

[flagged]


Since rafram is not the only one confused, yes, you really do.

It isn't that hard to understand:

> Just wait until there are entire classes of vulnerabilities related to LLM usage

This is a valid concern.

There are going to be a new class of vulnerabilities which an LLM is involved which are going to be discovered and it will make it possible to cause catastrophic damage to a company; very easily.

This won't be surprising since we have companies building casual remote code execution tools for "agents" waiting to be hijacked.


I understand that. What about that relates specifically to the Android CLI? That was rafram's question, and mine, and as far as I can tell still hasn't been answered.

I mean, I guess if you're going to say "don't use LLMs", then you also don't want to let agents use the Android CLI, but it seems like raising an awfully general concern in a discussion about a very specific article.


That probably depends on how good 2026-era LLMs already are. But I hope you’re right, and that pre-AI devs will still make a real difference.

Catching up to Flutter.

Not even close. Flutter has been engineered from the ground up with excellent tooling, unlike Android’s mess of organically evolved crap held together by a duct tape.

Yeah, I've been using Flutter since December last year and I'm really amazed how good the developer experience is. I kinda regret not picking it up sooner but from what I understand now's a great time too with the roadmap they've planned for this year (videos on YouTube explaining Flutter concepts and decoupling Material and Cupertino).

You can even make 2D games with Flutter with the help of Flame[0] but be wary that pixel art style games are a bit of a hassle due to some bugs in Flutter itself. Otherwise Flutter is a joy to use for its intended purpose: cross-platform apps.

[0] https://flame-engine.org/


flutter have this already?

AFAIK, Flutter has had a good, capable CLI since the beggining. You've never needed to install Android Studio to use Flutter.

I meant in terms of development speed with agents.

Flutter CLI is what we really need but this is a welcome addition.

It exists already. Wdym?

Flutter quality like tooling, I assume.



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: