Android Feed

← The 5 stories that mattered

Everything collected ·

All 344 items collected for this issue

New here? How to read this issue
  1. Read the picks. Each tab opens with 4–5 stories ranked by how much you’d regret missing them. That takes about 3 minutes. “Start the read” shows one story at a time, and “Listen” reads them aloud.
  2. Stop at “You’re caught up”, or go deeper. Everything after it is the full list of what was collected today, folded away.
  3. Android comes first. Product & AI covers what builders are launching and discussing.
Stable · Beta · Alpha
Release stages. Stable is safe to ship, beta’s API is frozen but may still have bugs, alpha is experimental.
AndroidX
Google’s Jetpack libraries (Compose, Paging, Room…), released separately from Android itself.
AOSP insider
Code changes merged into Android’s open source repository. They often show an API before it’s announced.
Lanes
Where an Android item comes from: official, AOSP insider, experts, open-source library releases, leaks and ecosystem, community.
Adopt now · Watch · Rumour
What to do about it, and how sure the source is.
Dotted underline
An inferred claim: Claude’s reading, not something the source says.

Collected automatically each morning from 37 sources. Notes are written by Claude. Links show what was trending, not endorsements.

Android5Product & AI4
Start the Android read5 stories, ranked by what you’d regret missing · 2 min read

  1. New Jetpack Appstate libraries bring Compose state beyond the UI

    Compose's state model is moving past screens, which could change how you write ViewModels and app-wide state.

    Google released the first alphas of two new Jetpack libraries on October 28. Appstate Statestore lets apps move snapshot state, the observable state Jetpack Compose uses, into the app layer, with an optional link to DataStore for saving it. Appstate Transform runs a composition outside the UI, so logic can be written in the Compose style and tested. Both are early alphas.

    AndroidX releases · OfficialRead the sourceProduct notes

  2. LiteRT in Play services gets faster GPU inference

    If your app runs models on device, switching to the Play services runtime makes it smaller and faster at once.

    Google's Developers Blog says LiteRT, its runtime for running machine learning models on the device, now offers GPU inference inside Google Play services. It is powered by a new engine called ML Drift and is about 1.4 times faster, with a new CompiledModel API. Because the runtime comes from Play services, apps no longer need to bundle it.

    Google Developers Blog · OfficialRead the sourceProduct notes

  3. One malformed text message can crash Android's phone service

    Every app that relies on calls or texts depends on this shared service, so watch for the fix in upcoming security patches.

    A report shared on Hacker News and Lobsters says a single malformed WAP Push message, a kind of system text message, crashes com.android.phone, the process that handles calls. When it crashes, active calls drop, including emergency calls. The Hacker News thread has about 330 points and nearly 190 comments.

    Lobsters · android · CommunityRead the sourceDiscussion

  4. Did AI kill cross-platform mobile development?

    If coding agents make writing native code twice cheap, the case for one shared codebase gets weaker.

    Vasiliy Zukanov of TechYourChance looks at Shopify's move away from React Native, Meta's cross-platform framework. Shopify claims AI has changed the trade-off between sharing one codebase and building native apps for Android and iOS. Zukanov reviews the evidence behind that claim.

    TechYourChance (Vasiliy Zukanov) · ExpertRead the source

  5. Haze 2.0 adds glass effects to Compose Multiplatform

    A major version means API changes, so read the notes before upgrading if your app uses Haze.

    Chris Banes released Haze 2.0, his library for blur and frosted glass effects in Compose Multiplatform. It adds a Glass effect, a new effects API, performance improvements and experimental support for Android's native backdrop effects. These are the see-through bars and sheets that Apple's Liquid Glass design made popular again.

    Chris Banes · ExpertRead the source

Go deeper: every Android item, by lane

Official & releases

17Look at the new Appstate Statestore and Transform alphas, which bring Compose's state model outside the UI, and switch to LiteRT in Play services for faster GPU inference; Media3 1.12 and Test Core 1.8 are still alphas, and Studio Rabbit 2 remains the stable pick.

AndroidX library releases

Alpha · experimental, APIs may change · 4
Show 5 more

AOSP / AndroidX insider

10Android XR's scene code now batches changes and drops manual transactions, Remote Compose hides its internals and gains a surface modifier for glasses UI, and a WebView change points to sharing memory buffers between Java and web content.
  • CameraX: default check for extension support

    Add default implementation for CameraMetadata#isExtensionSupported

    A default implementation means fakes and custom metadata classes no longer have to override it, API polish while Camera 1.7 is at release candidate.

  • Android XR: scene changes batched together

    Add a coalescing committer for OpenXR scene graph changes

    A coalescing committer groups many scene graph edits into fewer OpenXR commits, performance work for SceneCore now that it is 1.0.

  • Android XR: manual scene transactions removed

    Remove OpenXrTransaction and cancelSceneTransaction

    Paired with the batching change, apps no longer open or cancel scene transactions by hand; the runtime manages them.

  • Feathered edges: equality and a zero-size guard

    Add feathering effect equality and divide by 0 protection

    The soft-edge feathering effect gets equality checks, which avoid needless updates, and a divide-by-zero guard against crashes at zero size.

  • Glasses UI: surface modifier for remote layouts

    RC Glimmer: Add RemoteModifier.surface API

    Remote Compose meets Glimmer, the design system for Android XR display glasses, a sign that glasses UI may be drawn remotely from a phone app.

  • Compose Remote: draw paths from raw arrays

    Add addPathData(float[], int) overload to RemoteComposeWriter

    A lower-level path writer keeps vector shapes compact in remote UI documents.

  • Dialogs: tunable gesture thresholds

    Add gesture thresholds to AlertDialogDefaults

    Gesture thresholds in AlertDialogDefaults suggest dialogs that respond to swipes or back gestures, with limits apps can tune.

  • Compose Remote: internals hidden from public API

    RCA-009: Hide creation and core internals from public API

    Creation and core internals leave the public surface, the kind of trimming done before a beta.

Show 2 more
  • WebView: share memory buffers with Java

    WebView Java SharedArrayBuffer API

    AndroidX WebKit is adding a Java API for SharedArrayBuffer, so apps could pass large binary data between native code and web content without copying.

  • CameraX: Ultra HDR sizes checked only on Android 15+

    Only query JPEG_R extension sizes on API 35+ in Camera2ExtensionsVendorExtender

    JPEG_R is the Ultra HDR format; skipping the query on older versions avoids failures with vendor camera extensions.

Experts

18Read TechYourChance on whether AI killed cross-platform development, Chris Banes on Haze 2.0 and on when agent tasks need a plan, and skydoves' three-part Compose internals series.
Show 10 more

Library releases

11kotlinx.serialization 1.12.0 is stable on Kotlin 2.4.10 and safe to adopt, KSP 2.3.13 adds context parameter support, OkHttp 5.5.1 and 4.12.1 are patches worth taking alongside this month's OkHttp CVE advice, Metro 1.5.x, Ktor 3.6.1, Apollo 4.4.4 and Roborazzi 1.77.0 are safe updates, and Decompose 3.6.0 is still beta.
  • Compose Unstyled 2.10.2

    Patch Changes 649c471 : Make dialogs, modal bottom sheets, and modal drawers behave inside of a Host the same as without. Tab and Shift+Tab now stay inside the topmost modal, background controls receive no keystrokes, and closing the modal restores the previous focus. b59481b …

  • Metro 1.5.1

    Fixes [IR/IC] Avoid recompiling unrelated files in the graph module when the constructor of an upstream @Contributes* class changes. [Gradle] Fix false hidden-dependency reports for contributions from associated compilations, such as main sources in tests. This particularly…

  • Metro 1.5.0

    New [metrox-viewmodel-compose] Add support for linuxX64 . [FIR] Add the diagnostic-level compiler option to change the severity of Metro diagnostics reported in FIR. It works like kotlinc's -Xwarning-level but can also lower errors and use IDE-only levels. See changing…

  • OkHttp 5.5.1

    Version 5.5.1

  • OkHttp 4.12.1

    Version 4.12.1

  • KSP 2.3.13

    What's Changed Add KSContextParameter and KSFunctionDeclaration.contextParameters to the API ( #3216 ) Improve diagnostics by returning last round dismissible errors when KSP fails ( #3196 ) Add opt-in processor classloader caching ( ksp.classloader.cache.processors=true ) to…

  • kotlinx.serialization 1.12.0

    This release is based on Kotlin 2.4.10 and contains all of the changes from 1.12.0-RC . Additionally, it includes a fix removing unnecessary leading whitespaces from Proguard rules ( #3270 ), and two changes stabilizing kotlin.uuid.Uuid serializer ( #3263 ) and…

  • Ktor 3.6.1

    Published 29 October 2026 Improvements KTOR-10005 Native accepted TCP sockets do not enable TCP_NODELAY, unlike JVM KTOR-9966 Netty poor performance on delay calls KTOR-4206 The "com.typesafe.config.ConfigException$Missing" exception leaks from the typesafe config library when…

Show 3 more
  • Apollo Kotlin 4.4.4

    A patch version to fix Java codegen generating bad code when using custom scalars mapped to Long or Float. 👷‍♂️ All changes [FIX] codegen: fix Java codegen generating bad code when using custom scalars mapped to Long or Float. ( #7032 )

  • Decompose 3.6.0-beta02

    What's Changed Library Add linuxX64 and linuxArm64 targets by @arkivanov-bot in #1028 Fix wrong directions in AndroidPredictiveBackAnimatableV1 in RTL layout by @arkivanov in #1030 Fix wrong directions in AndroidPredictiveBackAnimatableV2 in RTL layout by @arkivanov in #1031 Add…

  • Roborazzi 1.77.0

    Bugfix: verifyRoborazzi* and compareRoborazzi* reuse the configuration cache With the configuration cache enabled, the second verifyRoborazzi* or compareRoborazzi* run after clean recalculated the task graph because "the file system entry 'build/outputs/roborazzi' has been…

Leaks & ecosystem

21Teardowns point to an offline mode and a new look for Gemini Live, Android 17 sits near 1 percent of devices, and Nothing and Samsung keep rolling out Android 17 updates.
Show 13 more

Community

37The malformed SMS that crashes Android's phone service leads Hacker News Android discussion, Shizuku and v2rayNG top Kotlin trending, and ProAndroidDev publishes a 2027 Android developer roadmap.
Show 29 more
Start the Product & AI read4 stories, ranked by what you’d regret missing · 2 min read

  1. A shared ChatGPT link was enough to take over accounts

    A single loose message listener in a third-party library can expose every logged-in user, so audit yours.

    Researchers published how they took over ChatGPT accounts, and the write-up reached about 300 points on Hacker News. Techmeme reports OpenAI paid them 5 thousand dollars for the bug. Opening a shared ChatGPT link let attackers run code in the viewer's logged-in session. A Dev.to post traces it to a math keyboard library that accepted messages from any website.

    HN front page · Across sourcesRead the sourceDiscussion

  2. HackerTube builds a video site from Hacker News comments

    Years of community recommendations sit in public archives, and turning them into a product is now a weekend job.

    HackerTube is a YouTube-style site made from videos that Hacker News readers recommended in comments since 2007. It covers about 6,500 videos found in nearly 40 thousand comments. It was the second story on Hacker News, with about 780 points and 250 comments.

    HN front page · Fastest risingRead the sourceDiscussionProduct notes

  3. claude-mem passes 100 thousand stars as agents get memory

    Agents that remember past sessions waste less time, and memory is becoming a layer separate from the agent itself.

    claude-mem records what a coding agent does in a session, compresses it with AI and feeds the useful parts back next time. It works across many agents, not just Claude Code. It gained about 1,700 stars in a day, the most on GitHub, and is near 108 thousand.

    GitHub Trending · Fastest risingRead the sourceProduct notes

  4. A language model squeezed into a 512-byte boot sector

    It is a playful reminder of how small a useful model can get, as builders push AI onto phones and tiny devices.

    A Show HN project fits a tiny language model into a computer's boot sector, just 512 bytes. The author took the idea from the demo scene, where people have written chess, BASIC games and even a C compiler in the same space. It appeared on Show HN and the Hacker News front page.

    Show HN · New ideaRead the sourceDiscussionProduct notes

Go deeper: themes, product notes, every source

Themes across sources

6Chat share-link exploits · Coding-agent control surfaces · Open-weight frontier models · On-device AI · Verifying agent work · Anti-slop design skills

Ideas showing up in more than one place today. The line counts items per day over two weeks.

Coding-agent control surfaces

Rising · 11 items from 2 sources

Orkas, a local-first desktop team where a Commander agent splits a goal into to-dos for 16 specialist agents, added about 1,000 stars. t3code, OpenMausBot and Claude Code Game Studios keep climbing on GitHub. Product Hunt adds GitHub Copilot Fusion, which picks and combines models, SpaceO for giving agents their own Mac screen, AC Bridge for smart lights that signal when an agent needs you, and MAO, ilse and Faylee return.

Why now. People now run several agents at once, so the product questions are how to assign them work, where they run and how to know when they need a human.

Orkas-AI/OrkasGitHub Trending

More on this

Open-weight frontier models

Rising · 6 items from 2 sources

Memdoor claims a team of five Qwen3.8-27B models beats Fable 5 on LiveCodeBench Hard, drawing about 95 points on Hacker News. On Hugging Face, Nanbeige 4.3 33B, DeepSeek V4.1 Flash, Mistral Large 4 and Qwen3.8-27B fill the top of the trending list, with a compressed DeepSeek build for local use.

Why now. Mid-sized open models are now good enough that orchestrating several of them is a credible alternative to one closed frontier model.

Memdoor: A team of 5 Qwen3.8-27B models beats Fable 5 on LiveCodeBench HardHN front page

On-device AI

Rising · 9 items from 8 sources

Teardowns show Google working on an offline mode for Gemini Live, covered by Android Authority and its Insights team. LiteRT in Google Play services gains faster GPU inference, Liquid AI's LFM3 small models trend on Hugging Face with a browser demo, and a Show HN fits a language model into a 512-byte boot sector. Sam Altman argues scarce memory is better used in data centers than in PCs.

Why now. Small models now handle everyday tasks, and runtimes on phones and in browsers have matured enough to ship them.

No internet, no problem: Google is working on an offline mode for Gemini LiveAuthority Insights (leaks & teardowns)

Verifying agent work

Rising · 7 items from 4 sources

Y Combinator launches Hue, which tests agents in stateful clones of the apps they work in, Composal, a swarm of agents that tests products like real users, and Mirrors, which corrects repeated agent mistakes. Product Hunt lists telemetry.dev for agent observability and Specsight for product observability, and posts on Dev.to and ProAndroidDev ask how to debug code an agent wrote.

Why now. Agents now act in real apps and write production code, so checking their work has become the bottleneck rather than generating it.

Hue: Simulations for AI agentsYC Launches

Anti-slop design skills

Rising · 4 items from 2 sources

Impeccable, a design vocabulary for coding agents, is near 98 thousand stars, Superpowers passed 310 thousand and Matt Pocock's skills collection is near 308 thousand. Dev.to adds a guide to measuring how agent skills drift out of date while their checks stay green.

Why now. Skills are a shared file format across most coding agents, so rules written once reach every agent a team uses, and keeping them current is the new chore.

pbakaus/impeccableGitHub Trending

Fading: Agent memory layers (last seen 2026-10-31) · Agent supply-chain safety (last seen 2026-10-31) · World models (last seen 2026-10-31) · Agent-ready Android (last seen 2026-10-29) · Android XR glasses (last seen 2026-10-29) · Agents as security auditors (last seen 2026-10-26) · Open voice models (last seen 2026-10-26) · Agent sandboxes (last seen 2026-10-25)

Product notes

15Who each product is for, the problem, why now, its wedge and how it makes money. Today’s picks first.

Plain text: the source says soDotted: inferred

HackerTube: YouTube built from videos recommended by HN

discussionHN front page782 pts · 251 comments

What
HackerTube, a YouTube-style site over 6,454 videos drawn from 39,834 Hacker News comments since 2007.
For whom
Hacker News readers who want talks and videos the community recommended, without YouTube's feed.
Problem
Good video recommendations are buried in old comment threads, and YouTube's algorithm favors engagement.
Why now
Models make it cheap to mine years of comments for links and context.
Wedge
A trusted community's taste as the recommendation engine.
Money
Side project, no model.
Like
Hacker News · YouTube · Nebula · Feedly

LessonCuration by a trusted crowd can beat an algorithm; the data may already exist in public archives.

thedotmack/claude-mem

repoGitHub Trending★ 1,723 today

What
claude-mem captures what an agent does in a session, compresses it with AI and injects relevant context back into later sessions, across many coding agents.
For whom
Developers who use coding agents daily on the same projects.
Problem
Each new session forgets earlier decisions, so people keep long notes files or re-explain the project.
Why now
Agent hooks and long sessions are common, so there is a record worth saving and a place to inject it.
Wedge
Works with every agent instead of locking memory to one vendor.
Money
Open source, no model yet.
Like
Mem0 · Letta · Zep · CLAUDE.md files

LessonNeutral glue between tools that do not talk to each other gets adopted faster than another walled garden.

Show HN: A language model that lives in a boot sector

productShow HN, HN front page60 pts · 1 comments

What
A Show HN project that fits a tiny language model into a 512-byte boot sector, inspired by the demo scene's chess and BASIC games in the same space.
For whom
Low-level programmers and anyone curious how small a language model can get.
Problem
None practical; it tests the limits of what a model needs.
Why now
Interest in small models is high, and the demo scene gives a famous constraint to compete against.
Wedge
An extreme, easy-to-share constraint.
Money
Hobby project, no model.
Like
sectorlisp · BootChess · llama2.c

LessonA hard, famous constraint makes a project memorable; size limits are a story by themselves.

Appstate Statestore 1.0.0-alpha01

releaseAndroidX releasesAndroid

What
Appstate Statestore 1.0.0-alpha01, a new Jetpack library built on the Compose runtime that lets apps hoist snapshot state into the app layer, with an optional DataStore link for saving it. Its sibling, Appstate Transform, runs a composition outside the UI.
For whom
Android teams that already think in Compose state and want the same model for repositories and app-wide state.
Problem
App state lives in a mix of Flows, LiveData and hand-written stores, and syncing it to disk takes glue code.
Why now
The Compose runtime is now mature and multiplatform, so Google can reuse it as a general state engine, not only for UI.
Wedge
A wrapper for custom integrations plus a ready DataStore sync, so teams can try it on one piece of state.
Money
Free Jetpack library.
Like
Molecule · Circuit · Redux-style stores · Kotlin StateFlow

LessonReuse an engine your users already trust in a new place; adoption costs less than teaching a new model.

Smaller apps and faster GPU inference with LiteRT in Google Play services

articleGoogle Developers BlogAndroid

What
LiteRT, Google's on-device machine learning runtime, now offers GPU inference in Google Play services powered by ML Drift, about 1.4 times faster, with a new CompiledModel API.
For whom
Android apps that run models on device for vision, audio or text and want smaller downloads.
Problem
Bundling an ML runtime and GPU delegates adds megabytes to every app and varies across devices.
Why now
More apps ship on-device models, so a shared runtime delivered by Play services saves space across the whole phone.
Wedge
Smaller apps for free when you switch to the Play services runtime.
Money
Free; Google gains by making Android the easiest place to run on-device AI.
Like
Core ML · ONNX Runtime Mobile · ExecuTorch · MediaPipe

LessonLead with the benefit users feel, smaller apps, and let speed be the second reason to switch.

agupta07505/SmartIsland

repoGitHub Trending · KotlinAndroid★ 12 today

What
Smart Island, an open-source take on a Dynamic Island for Android that turns notifications, media, calls and timers into a floating pill.
For whom
Android users who like the iPhone's Dynamic Island, and developers studying overlay and notification APIs.
Problem
Android has no system-wide live activity pill, so users rely on paid launcher add-ons.
Why now
Live updates and media sessions are richer on recent Android versions, making a polished overlay feasible.
Wedge
A familiar iPhone pattern, free and open source.
Money
Open source, no model yet.
Like
dynamicSpot · Samsung Now Bar · iOS Dynamic Island

LessonCopying a well-known pattern to a platform that lacks it is a fast way to find early users.

Orkas-AI/Orkas

repoGitHub Trending★ 1,044 todayHype flag

What
Orkas, an open-source, local-first AI team for the desktop: a Commander turns a goal into to-dos and coordinates 16 built-in specialist agents plus Claude and others.
For whom
Individuals who want several agents on one goal without running a cloud service.
Problem
Running many agents means juggling terminals and splitting work by hand.
Why now
Single agents are reliable enough that coordination between them is the next friction.
Wedge
MIT license and local-first, with a manager agent doing the planning.
Money
Open source, no model yet.
Like
CrewAI · MAO · Claude Code Game Studios · AutoGen

Sixteen specialist agents is a feature count, not evidence that the team finishes tasks better than one agent.

LessonBorrow an org chart people understand, a lead and specialists, to make a multi-agent product legible.

Memdoor: A team of 5 Qwen3.8-27B models beats Fable 5 on LiveCodeBench Hard

discussionHN front page95 pts · 46 commentsHype flag

What
Memdoor, a setup where a team of five Qwen3.8-27B open models works together and, by its own claim, beats Fable 5 on LiveCodeBench Hard.
For whom
Teams that want frontier-level coding results from models they can host.
Problem
The best coding models are closed and priced per token, and single open models lag behind them.
Why now
Mid-sized open models are cheap enough to run several at once on one server.
Wedge
A benchmark win that is easy to state and check.
Money
Not stated.
Like
Mixture-of-agents · DeepSeek V4.1 · Qwen Code · Fable 5

One benchmark result; cost, latency and real-project performance are not shown in the brief.

LessonA clear head-to-head claim spreads fast, but buyers will ask what it costs per task.

Show HN: A USB-C LTE modem that gives my computer a phone line

productShow HN47 pts · 34 comments

What
A USB-C LTE modem that gives a computer its own phone line, shown on Show HN.
For whom
People who want to send texts or take calls from a computer without a phone.
Problem
Computer calling and texting usually depend on a paired phone or an internet calling service.
Why now
Cheap LTE modules and USB-C make a small, self-contained modem practical for hobbyists.
Wedge
One plug-in device instead of a phone or service subscription.
Money
Not stated; possibly a hardware sale later.
Like
Phone Link · Twilio · USB LTE dongles

LessonGive a familiar capability a new home; a phone line in a laptop reframes what the laptop is for.

Hue: Simulations for AI agents

productYC Launches66 votes

What
Hue, a Y Combinator company that tests AI agents in stateful clones of the apps they work in.
For whom
Teams building agents that act inside SaaS apps like CRMs or ticket systems.
Problem
Testing agents against real apps is risky and hard to reset, so teams test by hand or in production.
Why now
Agents now take real actions, and a mistake in a live account costs money or trust.
Wedge
Clones that keep state, so a test run behaves like the real app.
Money
Likely usage-based SaaS.
Like
Composal · Braintrust · Browserbase · staging environments

LessonEvery new kind of software needs its own staging environment; building it is a business.

Aviern: Unmanned cargo airplanes for autonomous airports

productYC Launches188 votesHype flag

What
Aviern builds unmanned fixed-wing cargo airplanes for regional freight and defense logistics, claiming 1,000 pounds carried over 1,000 miles for under 1,000 dollars.
For whom
Freight and logistics operators serving regional routes, and defense buyers.
Problem
Small regional cargo flights are expensive because of pilots and airport costs.
Why now
Autonomy and drone regulation have advanced enough to plan pilotless cargo routes.
Wedge
Cargo only, between autonomous airports, avoiding passenger certification.
Money
Likely freight-as-a-service or aircraft sales.
Like
Reliable Robotics · Xwing · Elroy Air · Natilus

The 1,000-1,000-1,000 cost claim is a pitch number; certification and real operating data are not in the brief.

LessonA memorable three-number claim can carry a whole pitch; make sure it survives scrutiny.

GitHub Copilot Fusion

productProduct Hunt#17 on Product Hunt

What
GitHub Copilot Fusion, GitHub's AI coding agent that picks and combines models for a task.
For whom
Developers on GitHub who do not want to choose a model per task.
Problem
Model choice changes weekly, and users guess which one fits each job.
Why now
Several frontier models are close in quality but differ by task, so routing pays off.
Wedge
Built into GitHub, where the code already lives.
Money
Part of paid Copilot plans.
Like
Cursor Auto · Intrascope · OpenRouter · Claude Code

LessonWhen options multiply, choosing for the user becomes the product.

Knucklebop

productProduct Hunt#2 on Product Hunt

What
Knucklebop lets you tap your Mac's palm rest to control it.
For whom
Mac users who want quick shortcuts without reaching for keys or the trackpad.
Problem
Common actions need key combos that are hard to remember.
Why now
Laptop motion sensors and trackpad haptics make tap detection on the case possible.
Wedge
A new gesture on hardware users already own.
Money
Likely a paid Mac app.
Like
BetterTouchTool · Raycast · Back Tap on iPhone

LessonFind an unused input on existing hardware; it costs users nothing to try.

Godwit

productProduct Hunt#5 on Product Hunt

What
Godwit, an open-source, AI-native email client.
For whom
People who want AI help with email but do not want to hand their inbox to a closed service.
Problem
AI email tools are mostly closed and read all your mail on their servers.
Why now
Capable models can triage and draft mail, and local or bring-your-own-key setups are common.
Wedge
Open source as the privacy guarantee.
Money
Open source; a hosted plan is likely.
Like
Superhuman · Shortwave · Scape · Thunderbird

LessonIn a category of closed AI tools, openness itself is a wedge for privacy-minded users.

LiquidAI/LFM3-Omni-2B

modelHF trending modelstrending 231

What
LFM3-Omni-2B, a small multimodal model from Liquid AI, trending on Hugging Face with a WebGPU demo that runs in the browser.
For whom
App developers who want speech, vision and text in one model on phones, laptops or the web.
Problem
Multimodal features usually need a cloud call, adding cost, delay and privacy questions.
Why now
Small models and WebGPU have improved enough to run multimodal tasks locally.
Wedge
A 2 billion parameter size picked for edge devices.
Money
Open weights with licensing and enterprise deals.
Like
Gemma · Phi · Qwen Omni · Gemini Nano

LessonA demo that runs in the browser removes every excuse not to try your model.

By source

11Everything collected today from GitHub Trending, Product Hunt, Show HN, HN front page, YC Launches, HF trending models, HF trending spaces, HF daily papers, Lobsters, Dev.to top, Techmeme, in each source’s own order.

GitHub Trending

14
  • thedotmack/claude-memcard

    Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More

    ★ 1,723 today

  • mattpocock/skills

    Matt Pocock's skills collection adds about 880 stars, near 308 thousand.

    ★ 883 today

  • Donchitos/Claude-Code-Game-Studiosnew

    Turns Claude Code into a game studio with 49 agents and 72 skills arranged like a real studio hierarchy.

    ★ 742 today

  • pingdotgg/t3code

    t3code adds over 1,100 stars, one of the busiest coding-agent repos today.

    ★ 1,148 today

  • Renhuai123/ziwei-doushu

    An open-source Zi Wei Dou Shu astrology chart engine, among today's fastest risers.

    ★ 1,092 today

  • milind-soni/OpenMausBot

    OpenMausBot, an open-source bot with its own virtual machine, adds about 1,300 stars.

    ★ 1,303 today

  • pbakaus/impeccable

    Impeccable, a design vocabulary for coding agents, adds about 1,300 stars and nears 98 thousand.

    ★ 1,338 today

  • obra/superpowers

    Superpowers, the agentic skills framework, passes 310 thousand stars.

    ★ 665 today

Show 6 more
  • lightpanda-io/browser

    Lightpanda, a headless browser built for agents rather than people, adds about 900 stars.

    ★ 929 today

  • HKUDS/DeepTutor

    DeepTutor, an open-source personal tutor, keeps climbing past 52 thousand stars.

    ★ 1,034 today

  • Orkas-AI/Orkasnewcard

    Orkas is an open-source (MIT), local-first AI team for your desktop: a Commander turns your goal into to-dos and coordinates 16 built-in specialist agents plus Claude Code, Codex, OpenCode, OpenClaw and Hermes, with projects, scheduled automations and connectors. Free app for…

    ★ 1,044 today

  • THU-MAIC/OpenMAICnew

    OpenMAIC, an open multi-agent interactive classroom, past 42 thousand stars.

    ★ 116 today

  • bregman-arie/devops-exercisesnew

    A long-running set of DevOps interview exercises, trending again.

    ★ 342 today

  • vorssaint/vorssaint-utilsnew

    A free, open-source macOS menu bar toolkit.

    ★ 501 today

Product Hunt

50
  • Hemmednew

    Hemmed hides Mac menu bar clutter and pins single actions, the top Product Hunt launch.

    #1 on Product Hunt

  • Knucklebopnewcard

    tap your palm rest to control your Mac

    #2 on Product Hunt

  • ChainPointsnew

    Shows flight prices in loyalty points directly on Google Flights.

    #3 on Product Hunt

  • WinnerAdSpynew

    Turns competitor ad insights into app install campaigns.

    #4 on Product Hunt

  • Godwitnewcard

    An open-source, AI-native email client

    #5 on Product Hunt

  • telemetry.devnew

    telemetry.dev offers observability for AI agents.

    #6 on Product Hunt

  • BaseTab AInew

    An AI workspace in your browser's new tab.

    #7 on Product Hunt

  • Seqoranew

    Design a database schema on a canvas and get the SQL written for you.

    #8 on Product Hunt

Show 42 more
  • Backenlynew

    Backenly pitches a backend that coding agents can run on their own.

    #9 on Product Hunt

  • Agnosnew

    Cleans documents for AI use and shows what was removed.

    #10 on Product Hunt

  • LumaFold: Screen Lightnew

    A screen-light makeup mirror for iPhone.

    #11 on Product Hunt

  • AC Bridgenew

    Smart lights that show when Claude Code or Codex needs your attention.

    #12 on Product Hunt

  • SpaceOnew

    SpaceO gives AI agents their own screen on your Mac so they do not take over yours.

    #13 on Product Hunt

  • Specsightnew

    Specsight tracks product behavior for teams shipping fast with AI.

    #14 on Product Hunt

  • Pluto Doornew

    An SSH client with a built-in agent and a desktop per server.

    #15 on Product Hunt

  • CubbyPadnew

    Brings back a Launchpad-style app grid on macOS 26.

    #16 on Product Hunt

  • GitHub Copilot Fusionnewcard

    GitHub's AI coding agent that picks and combines models

    #17 on Product Hunt

  • Secret Keynew

    Lets your family get your passwords if something happens to you.

    #18 on Product Hunt

  • Coding Wrappednew

    A Wrapped-style recap of how you use Claude Code and Codex.

    #19 on Product Hunt

  • AwaitIOnew

    Write a TypeScript function and get an admin screen for it.

    #20 on Product Hunt

  • SKInew

    SKI, a voice-first coding harness.

    #21 on Product Hunt

  • Monew

    A character on your Mac that claims to handle 232 tasks.

    #22 on Product Hunt

  • StereoLift Live

    Shows your desktop in 3D on a Quest headset.

    #23 on Product Hunt

  • Knocket AI

    A chat widget for talking to website visitors.

    #24 on Product Hunt

  • Norah Studio

    A new app for exploring your music library.

    #25 on Product Hunt

  • Nocto

    Another app that puts tools in the MacBook notch.

    #26 on Product Hunt

  • Chopstick

    Mac widgets that write themselves from your activity.

    #27 on Product Hunt

  • SpeakON Agent

    SpeakON turns spoken requests into tasks for an agent.

    #28 on Product Hunt

  • Kestrix

    Kestrix runs CI across repositories with dependency-graph triggers.

    #29 on Product Hunt

  • Cluinn

    Cluinn answers questions during a call on a Mac, not after it.

    #30 on Product Hunt

  • SparkBots AI (Powered by Jev)

    Extracts key details from any webpage or PDF.

    #31 on Product Hunt

  • StepGrab

    StepGrab's step-by-step guide maker reaches Windows with redaction.

    #32 on Product Hunt

  • Canva Brush

    Canva Brush, a vision-language model that edits your existing designs.

    #33 on Product Hunt

  • iSurge

    An activity monitor, per-app volume and disk cleanup in the Mac notch.

    #34 on Product Hunt

  • Teable Mobile

    Teable's no-code database builder comes to phones.

    #35 on Product Hunt

  • MCP Server by XWiki

    XWiki's MCP server gives AI assistants a wiki's knowledge while keeping permissions in place.

    #36 on Product Hunt

  • Vaani Developer Platform

    Vaani pitches a developer platform for building voice agents.

    #37 on Product Hunt

  • ilse

    ilse lets you point at your running app so your agent changes the matching code.

    #38 on Product Hunt

  • fastbrowse

    fastbrowse claims to be the fastest open-source browser agent.

    #39 on Product Hunt

  • sitesolide

    Deploy vibe-coded apps on your own server instead of a hosted platform.

    #40 on Product Hunt

  • Faylee

    Faylee turns a bug report into one link a coding agent can fix from.

    #41 on Product Hunt

  • Axonnic

    Muscle-sensing sleeves that record hand data for robot training, also a YC launch.

    #42 on Product Hunt

  • Enclave

    Enclave puts the context you need behind a single hover.

    #43 on Product Hunt

  • Confetti

    Confetti designs hardware gadgets from plain English.

    #44 on Product Hunt

  • Scape

    An email inbox that sorts and acts on mail for you.

    #45 on Product Hunt

  • MAO

    MAO puts Claude Code, Codex and Grok on one shared canvas.

    #46 on Product Hunt

  • ChatInbox

    An inbox grouped by contact instead of by thread.

    #47 on Product Hunt

  • Reset Alerts

    Text alerts when your Codex usage resets, so paid quota is not wasted.

    #48 on Product Hunt

  • Intrascope 2.0

    Intrascope 2.0 adds an agent, 1,650 app connections and model routing.

    #49 on Product Hunt

  • Viggle JST-2

    Viggle JST-2 makes 3D scenes, characters and motion from one model.

    #50 on Product Hunt

Show HN

6

HN front page

30
Show 22 more

YC Launches

20
Show 12 more

HF trending models

19
Show 11 more

HF trending spaces

17
Show 9 more

HF daily papers

20
Show 12 more

Lobsters

25
Show 17 more

Dev.to top

20
Show 12 more

Techmeme

15
Show 7 more

About this issue

344 items from 77 sources, 11 of them seen in several sources. 3 sources failed to fetch.

Source health

77Which sources were fetched today, and how many items each returned.
SourceSectionItemsStatus
GitHub TrendingMain14Fetched
Product HuntMain50Fetched
Show HNMain6Fetched
HN front pageMain30Fetched
YC LaunchesMain20Fetched
HF trending modelsMain20Fetched
HF trending spacesMain20Fetched
HF daily papersMain20Fetched
LobstersMain25Fetched
Dev.to topMain20Fetched
TechmemeMain15Fetched
AndroidX releasesAndroid · Official & releases13Fetched
Android Developers BlogAndroid · Official & releases5Fetched
Android Studio releasesAndroid · Official & releases5Fetched
Kotlin blogAndroid · Official & releases1Fetched
Kotlin releasesAndroid · Official & releases0Fetched
Compose Multiplatform releasesAndroid · Official & releases1Fetched
Android Developers YouTubeAndroid · Official & releases0Failed
Kotlin on YouTube (JetBrains)Android · Official & releases0Failed
Google Developers BlogAndroid · Official & releases1Fetched
AndroidX merged API changesAndroid · AOSP / AndroidX insider10Fetched
Dove Letter (skydoves)Android · Experts5Fetched
skydoves on MediumAndroid · Experts0Fetched
Jake WhartonAndroid · Experts0Fetched
Zac SweersAndroid · Experts0Fetched
Chris BanesAndroid · Experts2Fetched
Ryan HarterAndroid · Experts0Fetched
Talking KotlinAndroid · Experts0Fetched
Joe BirchAndroid · Experts2Fetched
CommonsWare (Mark Murphy)Android · Experts3Fetched
TechYourChance (Vasiliy Zukanov)Android · Experts1Fetched
Kt. Academy (Marcin Moskała)Android · Experts2Fetched
Touchlab (Kotlin Multiplatform)Android · Experts3Fetched
Landscapist releasesAndroid · Library releases0Fetched
Balloon releasesAndroid · Library releases0Fetched
Sandwich releasesAndroid · Library releases0Fetched
Cloudy releasesAndroid · Library releases0Fetched
Compose Stability Analyzer releasesAndroid · Library releases0Fetched
Haze releasesAndroid · Library releases0Fetched
Compose Unstyled releasesAndroid · Library releases1Fetched
Coil releasesAndroid · Library releases0Fetched
Circuit releasesAndroid · Library releases0Fetched
Metro releasesAndroid · Library releases2Fetched
Molecule releasesAndroid · Library releases0Fetched
Turbine releasesAndroid · Library releases0Fetched
Paparazzi releasesAndroid · Library releases0Fetched
SQLDelight releasesAndroid · Library releases0Fetched
OkHttp releasesAndroid · Library releases2Fetched
Retrofit releasesAndroid · Library releases0Fetched
LeakCanary releasesAndroid · Library releases0Fetched
Koin releasesAndroid · Library releases0Fetched
Dagger / Hilt releasesAndroid · Library releases0Fetched
KSP releasesAndroid · Library releases1Fetched
kotlinx.coroutines releasesAndroid · Library releases0Fetched
kotlinx.serialization releasesAndroid · Library releases1Fetched
Ktor releasesAndroid · Library releases1Fetched
Apollo Kotlin releasesAndroid · Library releases1Fetched
Kermit releasesAndroid · Library releases0Fetched
Store releasesAndroid · Library releases0Fetched
Decompose releasesAndroid · Library releases1Fetched
Orbit MVI releasesAndroid · Library releases0Fetched
Roborazzi releasesAndroid · Library releases1Fetched
detekt releasesAndroid · Library releases0Fetched
ktlint releasesAndroid · Library releases0Fetched
MockK releasesAndroid · Library releases0Fetched
Lottie releasesAndroid · Library releases0Fetched
Authority Insights (leaks & teardowns)Android · Leaks & ecosystem7Fetched
9to5Google · AndroidAndroid · Leaks & ecosystem8Fetched
Android AuthorityAndroid · Leaks & ecosystem6Fetched
Android FaithfulAndroid · Leaks & ecosystem1Fetched
Android WeeklyAndroid · Community15Fetched
GitHub Trending · KotlinAndroid · Community15Fetched
Lobsters · androidAndroid · Community1Fetched
HN · AndroidAndroid · Community4Fetched
ProAndroidDevAndroid · Community5Fetched
Kotlin WeeklyAndroid · Community1Fetched
Philipp Lackner (YouTube)Android · Community0Failed
1 of 5

New Jetpack Appstate libraries bring Compose state beyond the UI

Compose's state model is moving past screens, which could change how you write ViewModels and app-wide state.

Google released the first alphas of two new Jetpack libraries on October 28. Appstate Statestore lets apps move snapshot state, the observable state Jetpack Compose uses, into the app layer, with an optional link to DataStore for saving it. Appstate Transform runs a composition outside the UI, so logic can be written in the Compose style and tested. Both are early alphas.

AndroidX releases · OfficialRead the source

1 of 4

A shared ChatGPT link was enough to take over accounts

A single loose message listener in a third-party library can expose every logged-in user, so audit yours.

Researchers published how they took over ChatGPT accounts, and the write-up reached about 300 points on Hacker News. Techmeme reports OpenAI paid them 5 thousand dollars for the bug. Opening a shared ChatGPT link let attackers run code in the viewer's logged-in session. A Dev.to post traces it to a math keyboard library that accepted messages from any website.

HN front page · Across sourcesRead the sourceDiscussion