SHUBHAM SINGH BISHT

Android Daily ·

Android's command line tool now drives real phones and guides coding agents

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 · 3 min listen
  1. #1Official

    Android's command line tool can now drive real devices and guide coding agents

    Why it matters Agents writing Android code can now test on real hardware and follow Google's own advice instead of outdated examples.

    The Android Developers Blog announced two additions to the Android command line tool. Device Streaming lets you, or your coding agent, control real phones in Google's lab as if they were plugged in. Android skills are official instruction packs that ground agents in current guidance, including a new one for Wear OS with Compose Material 3.

    Read the original · Android Developers BlogProduct notes

  2. #2Official

    Jetpack Compose 1.13 reaches beta

    Why it matters This is the window to test your app on the new version and report regressions before it becomes the default.

    Jetpack Compose, Android's UI toolkit, moved to version 1.13 beta this week across UI, runtime, foundation, animation and material, alongside Material 3 version 1.5 beta. Beta means the API surface is frozen, so what you see now is what ships in stable. AndroidX also shipped stable updates to Collection, Paging and Wear Compose in the same release batch.

    Read the original · AndroidX releases

  3. #3Official

    Google's Agent Development Kit for Kotlin reaches version 1.0

    Why it matters Agents inside Android apps are becoming an official pattern, and now there is a first party Kotlin toolkit for building them.

    Google released version 1.0 of its Agent Development Kit for Kotlin, with the same features as the Python and Java versions. Kotlin and Android teams can now build agents without switching languages. It arrives alongside Google's new guide to agent workflows inside Android apps and a case study from Instagram's messaging team.

    Read the original · Google Developers BlogProduct notes

  4. #4Official

    Instagram's messaging team cut AI agent costs by a third with Compose

    Why it matters How you structure UI code now affects what AI agents cost to run on it, and this is one of the first numbers from a large production app.

    A joint post from Meta and Google explains how the Instagram Direct team, which handles billions of messages, structured its Jetpack Compose interface so that AI agents can work with it more efficiently. The team reports that the token cost of each agent session fell by 33 percent. The post walks through the architecture behind that number.

    Read the original · Android Developers Blog

  5. #5Official

    The Kotlin Multiplatform plugin is dropping Swift editing

    Why it matters Teams that edit their iOS Swift code inside Android Studio will need Xcode or another editor for that part.

    JetBrains is removing Swift features from its Kotlin Multiplatform plugin, the tool for sharing Kotlin code between Android and iOS. Starting with IntelliJ IDEA 2026.3 and Android Studio Rabbit 2, the plugin will no longer edit, highlight or navigate Swift files, including jumps between Swift and Kotlin. The Kotlin side of shared code is not affected.

    Read the original · Kotlin blog

02

Also worth knowing

Releases, policy and leaks an Android engineer should know about this week.

03 · Reference

Go deeper: every Android item, by lane

Official & releases

51Four AndroidX libraries went stable and are safe to update (Collection, Paging, Wear Compose, Navigation Event), Compose 1.13 is in beta, and Android Studio Rabbit 1 is now stable.

AndroidX library releases

Stable · safe to ship · 4

Beta · API frozen, try it before it ships · 14

Alpha · experimental, APIs may change · 14
Show 11 more

AOSP / AndroidX insider

3Google is finishing a batch passkey API in Credential Manager, removing old Android XR panel APIs, and making a Compose accessibility behaviour the default.

Experts

10skydoves explains how Compose decides a class is unstable and what that costs, and Chris Banes asks whether Compose's new progressive blur replaces his Haze library.
Show 2 more

Library releases

0

Nothing new here today.

Leaks & ecosystem

20Leaks point to Photo Picker editing and gamepad-as-mouse input, and Samsung is rolling Android 17 out to Galaxy phones, so test your app on it.
Show 12 more

Community

35LineageOS 24.0 is out, and this week's Android Weekly leans on Kotlin Multiplatform: Swift export, publishing one library everywhere, and shared telemetry.
Show 27 more
Start the Product & AI read4 stories, ranked by what you’d regret missing · 2 min read · 2 min listen
  1. #1Across sources

    Microsoft ships a small model just for agent decisions

    Why it matters Agents make many small calls per task, so a cheap, consistent judge can matter more than a smarter general model.

    Microsoft released Microsoft Decision 1, a fast model that scores decisions inside AI agents and workflows, such as whether to approve, route or rank something. The first version is built on Qwen 3.5 with nine billion parameters, and Microsoft says it will move it onto its own and OpenAI's models. The same idea showed up today on Hacker News, at a Y Combinator startup and on Hugging Face.

    Read the original · Product HuntProduct notes

  2. #2Fastest rising

    An agent-powered reverse engineering tool gains 25 thousand stars in a day

    Why it matters Work that needed rare experts is getting cheap, which matters for security, modding and app protection, though the tool's claims are still unproven.

    An open source project called rea promises to reverse engineer anything with AI agents, from how an app behaves down to its native binaries. It gained about 25 thousand GitHub stars in a single day. On the same day, reports described hundreds of old games being decompiled and ported to the browser with help from Claude.

    Read the original · GitHub TrendingProduct notes

  3. #3New idea

    New Mac tools tell you which coding agent needs you

    Why it matters When developers supervise several agents at once, the scarce resource is attention, and that is becoming its own product category.

    Several launches today tackle the same new problem: keeping track of many coding agents at once. Farol is a Mac terminal that shows which agent is waiting for you. Littleguys puts Claude Code and Codex agents in the menu bar, and Halo uses the MacBook notch for agent status. None of them write code. They route your attention.

    Read the original · Product HuntProduct notes

  4. #4Fastest rising

    Google's EmbeddingGemma 2 tops Hugging Face and already runs in the browser

    Why it matters Good multimodal search no longer needs a server, which opens up on-device search features for mobile apps.

    EmbeddingGemma 2 is Google's new open embedding model. It turns text, images, audio and video into vectors for search and recommendations, under the Apache license. It is the top trending model on Hugging Face, and within days the community had compressed versions and a demo running fully in the browser with WebGPU.

    Read the original · HF trending modelsProduct notes

02 · Reference

Go deeper: themes, product notes, every source

Themes across sources

6Small decision models · AI-driven reverse engineering · Coding-agent control surfaces · On-device AI · Self-improving agents · Agent-ready Android

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

Small decision models

new7 items5 sources

Microsoft launched a fast decision-scoring model for agent workflows, a 'build your own decision model' post sits high on HN, a YC company sells a fraud/risk decision model at a fraction of frontier cost, an 11MB local decision model launched on Product Hunt, and a decision-model leaderboard is trending on Hugging Face.

Why now. Agent pipelines now make dozens of small yes/no and ranking calls per task, so per-decision cost, latency and consistency matter more than raw intelligence.

Microsoft-Decision-1Product Hunt

AI-driven reverse engineering

new5 items3 sources

An agent-powered reverse-engineering repo gained ~25k GitHub stars in a day, a PS5-to-PC porting tool is trending, Techmeme reports hundreds of old games decompiled and ported to browsers with Claude, and HN has several hand-done reverse-engineering and porting stories.

Why now. Coding models can now read disassembly and iterate on decompiled code, collapsing the cost of work that used to need rare experts.

morluto/reaGitHub Trending

Coding-agent control surfaces

new8 items4 sources

Several launches attack the same new pain: supervising multiple coding agents. A terminal that shows which agent needs you, menu-bar and notch status for Claude Code/Codex agents, an always-on Mac agent, a mod to swap models inside Claude Code, a design-skill pack for agents, and a paper on critics for long-horizon coding agents.

Why now. Running several agents in parallel became normal practice, so the bottleneck moved from writing code to routing human attention.

FarolProduct Hunt

On-device AI

new12 items8 sources

Google's EmbeddingGemma 2 tops HF and already runs in the browser via WebGPU and as GGUF; Product Hunt has an on-device 'AI computer', offline assistants for phones and local document-to-speech; GitHub has 70B inference on a 4GB GPU; Android devs are fine-tuning small models for on-device agents.

Why now. Small open models (3B-27B) crossed the quality bar for real tasks, and WebGPU, GGUF and phone NPUs make shipping them practical.

google/embeddinggemma-2HF trending models

Self-improving agents

new9 items4 sources

A YC company sells agents that get cheaper and better with each run, papers on RL-driven self-improvement, reflective rulebooks, self-improving robot code and agent memory compaction trend on HF, a chip startup pitches hardware for recursive self-improvement, and agent infrastructure (persistent per-agent VMs, agent runtime security) is raising money.

Why now. Agents now run the same tasks thousands of times, which makes learning from past runs both possible and economically valuable.

Grep AI: Self-improving agents that get smarter and cheaper at the same taskYC Launches

Agent-ready Android

new8 items5 sources

Google shipped Android CLI device streaming and 'Android skills' to ground agents, ADK for Kotlin hit 1.0, the official blog published an in-app agentic workflows guide and a Meta case study on AI-native Compose UI, a Developer Knowledge MCP server went public, and AppFunctions keeps moving through alpha.

Why now. Coding agents are mainstream in Android teams and agents inside apps are becoming a platform feature, so Google is making its docs, devices and Kotlin stack machine-usable.

Device Streaming and Android skills - available in Android CLIAndroid Developers Blog

Product notes

16Who 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

productProduct Hunt

Microsoft-Decision-1

#40 on Product Hunt

What
A fast model from Microsoft that scores decisions inside agents and workflows, initially trained on Qwen3.5-9B.
For whom
Teams building agent pipelines that need cheap, consistent judgments (route, approve, rank) at each step.
Problem
Calling a frontier model for every small routing or approval step is slow, expensive and inconsistent.
Why now
Agent workflows make many small judgments per task, so per-decision cost dominates; it matches Nadella's call for deterministic systems around powerful models.
Wedge
A narrow job done cheaply and fast, from a vendor enterprises already buy from.
Money
Likely metered through Azure; Microsoft says it will rebase the model on MAI, OpenAI and other models.
Like
Grayson (YC) · ej · LLM-as-judge reward models

LessonWhen a general capability gets used in a narrow, repetitive slot, a specialised small model can win on cost and consistency.

repoGitHub Trendinghype flag

morluto/rea

★ 25.6k today

What
An open-source tool that uses AI agents to reverse-engineer software, from observed app behaviour down to native binaries.
For whom
Security researchers, modders and engineers who need to understand closed software without source.
Problem
Reverse engineering is slow, expert-only work in tools like Ghidra or IDA; most developers cannot do it at all.
Why now
Coding models can now read disassembly and iterate; the same wave is decompiling old games at scale (see Techmeme today).
Wedge
One open repo with a sweeping promise ('reverse engineer anything'), riding agent interest to viral GitHub growth.
Money
None stated; open source. A hosted or enterprise security product is the obvious later path.
Like
Ghidra · IDA Pro · Binary Ninja

'Anything' is a very broad claim, and ~25k stars in a day reflects virality more than proven results on real binaries.

LessonAn audacious promise on a hot capability buys distribution; retention depends on whether it survives contact with real targets.

productProduct Hunt

Farol

#9 on Product Hunt

What
A macOS terminal that shows which of your coding agents needs your attention.
For whom
Developers running several coding agents in parallel.
Problem
With multiple agents running, it is easy to lose track of which one is blocked waiting for input.
Why now
Running several agents at once is becoming normal; Littleguys, Halo and Agibot launched around the same idea today.
Wedge
Owning the terminal, where coding agents already run, instead of adding a separate dashboard.
Money
Not stated; likely a paid Mac app or a pro tier.
Like
Littleguys · Warp · iTerm2

LessonWhen users shift from doing to supervising, the valuable UI is attention routing, not more output.

modelHF trending models

google/embeddinggemma-2

trending 1,525

What
Google's EmbeddingGemma 2: an Apache-2.0 multimodal embedding model covering text, images, audio and video.
For whom
Developers building search, RAG and recommendations, including on-device and in-browser.
Problem
Strong multimodal embeddings have mostly been API-only or too large to run locally.
Why now
Within days it is quantised (Unsloth GGUF) and running in the browser over WebGPU, showing small open embeddings can be deployed anywhere.
Wedge
Open licence plus Google's name, small enough for devices.
Money
Free; it strengthens the Gemma ecosystem and Google Cloud.
Like
OpenAI text-embedding-3 · Jina Embeddings · Nomic Embed

LessonOpen, small and multimodal is a distribution strategy: the community ships the ports for you.

productYC Launches

Maritime: Fully managed cloud for running AI agents at scale

19 votes

What
A managed cloud where every AI agent runs on its own isolated, persistent VM.
For whom
Companies running many long-lived agents that need state, tools and isolation.
Problem
Serverless functions and shared containers fit badly with agents that run for hours, keep state and execute untrusted code.
Why now
Agents moved from chat turns to long-running jobs with real tool access, which needs persistent and safe compute.
Wedge
One infrastructure primitive, a VM per agent, promised at 'millions of agents' scale.
Money
Usage-based compute pricing, like other cloud platforms.
Like
E2B · Modal · Fly Machines

LessonA new workload shape (long-lived, stateful, untrusted) can open space for new infrastructure even in a mature cloud market.

productShow HNHN front page

Show HN: BetterWispr – Free, open-source dictation for Mac

95 pts · 65 comments

What
A free, open-source dictation app for Mac.
For whom
Mac users who dictate messages, documents and especially prompts to AI tools all day.
Problem
The best dictation apps are subscriptions, and built-in dictation is noticeably weaker.
Why now
Local speech models now run fast and accurately on Apple silicon, so a free app can match paid quality.
Wedge
Free and open source against a paid incumbent the name points at directly.
Money
None; open source.
Like
Wispr Flow · Superwhisper · macOS Dictation

LessonNaming yourself 'Better X' is cheap, instant positioning when you are the free alternative to a known paid product.

productShow HN

Show HN: WallHop – 12ft.io is gone, so I built a replacement

264 pts · 114 comments

What
A free web tool that shows a readable article without the paywall pop-up, ads or clutter; no signup and no extension.
For whom
Readers hitting soft paywalls, especially on phones (it ships an iOS Shortcut).
Problem
12ft.io, the usual tool, was taken down in July 2025 after pressure from the News Media Alliance.
Why now
The incumbent's removal left a gap; it chains several bypass methods and falls back to archive.is so one blocked trick doesn't break it.
Wedge
A URL-prefix trick plus a share-sheet shortcut: zero install, works from a phone.
Money
None stated; free. Legal pressure, the thing that killed 12ft.io, is the main business risk.
Like
12ft.io · archive.today · everywall/ladder

LessonA takedown leaves ready-made demand behind, but whatever removed the incumbent will come for the replacement too; plan for it from day one.

productProduct Hunthype flag

Ghost Core

#15 on Product Hunt

What
A 'personal AI computer' that runs entirely on-device.
For whom
Privacy-minded users who want an assistant without sending data to the cloud.
Problem
Cloud assistants see all your data, and do-it-yourself local setups are fiddly.
Why now
Small open models (Qwen3.8, LiquidAI d1, Gemma) are now good enough to run useful assistants locally.
Wedge
Privacy as the headline: nothing leaves the device.
Money
Unclear from the listing; 'computer' may mean a hardware sale.
Like
KernelAI · Ollama · Rabbit R1

'Personal AI computer' is vague; the listing doesn't say whether it is hardware or software or what jobs it does well.

LessonOn-device wins when paired with a concrete job; privacy on its own rarely sells a product.

productYC Launches

Grep AI: Self-improving agents that get smarter and cheaper at the same task

10 votes

What
A platform to build an agent in minutes that gets smarter and cheaper every time it runs the same task.
For whom
Teams automating repeated, high-volume workflows.
Problem
Agents cost as much on the thousandth run as on the first, and do not learn from what worked.
Why now
Self-improvement research (MiMo-V2.6, Memento 3) and cheap distillation make 'learn the task, then hand it to a smaller model' practical.
Wedge
Cost that falls with usage, a pitch finance teams understand immediately.
Money
Likely usage-based pricing that captures part of the savings.
Like
OpenPipe · Predibase · DSPy

LessonA unit cost that improves with volume is a strong story in any market where spend scales with usage.

repoGitHub Trending

cathrynlavery/diagram-design

★ 1,083 today

What
A skill pack that makes coding agents (Claude Code, Codex, Copilot, Cursor, Factory Droid, Pi) produce editorial-quality diagrams as self-contained HTML/SVG, with 44 diagram types.
For whom
Developers and technical writers who want good-looking diagrams straight from their agent.
Problem
Agent-generated diagrams, often Mermaid, look generic.
Why now
Agents now load portable 'skills', so design taste can ship as a file that works in many tools.
Wedge
A strong, opinionated style ('no shadows, no Mermaid slop') that works across six agent tools.
Money
None; open source. It builds the author's reputation.
Like
Mermaid · Excalidraw · D2

LessonIn the agent era, taste can be packaged and distributed like a library.

repoGitHub Trending

bilawalsidhu/gods-eye-view

★ 872 today

What
A 'spy satellite simulator' in the browser on a photorealistic 3D globe, using real open data.
For whom
OSINT enthusiasts, educators and curious people.
Problem
Spatial-intelligence tools are expensive, closed and built for analysts, not the public.
Why now
Photorealistic 3D tiles and open live data feeds make this buildable as a single web app.
Wedge
A cinematic, shareable framing: a spy satellite, except the data is real.
Money
None; open source.
Like
Google Earth · CesiumJS · Flightradar24

LessonA vivid metaphor that names a fantasy turns a data viewer into something people share.

productProduct Hunt

Edgee Mod for Claude Code

#10 on Product Hunt

What
A mod that lets you use any model inside Claude Code while keeping your Claude plan.
For whom
Claude Code users who want other models for some tasks.
Problem
Developers like Claude Code's harness but want model choice, or cheaper models for routine work.
Why now
The agent harness has become the part people value; the model underneath is increasingly swappable.
Wedge
Plugging into a tool developers already use daily rather than asking them to switch.
Money
Likely gateway usage fees from Edgee, which runs an AI gateway.
Like
OpenRouter · LiteLLM · claude-code-router

LessonA mod for a tool people love is a fast route to distribution, but you live at the host's mercy.

productYC Launches

Invofox – Turn any document into structured, validated JSON

10 votes

What
A document-processing API: send any PDF or image, get accurate, validated structured JSON back.
For whom
Software teams that process invoices, forms and other documents.
Problem
Extraction is error-prone, and raw model output needs validation before it can go into a database.
Why now
Vision models made extraction cheap; the remaining value is accuracy and guarantees.
Wedge
Developer-first API with validation built in.
Money
Per-document API pricing.
Like
Reducto · AWS Textract · Unstructured

LessonWhen a capability becomes a commodity, sell the guarantee (validated output), not the capability.

articleAndroid Developers BlogAndroid

Device Streaming and Android skills - available in Android CLI

What
Android CLI adds Device Streaming (drive remote physical devices) and 'Android skills' that ground coding agents in official guidance.
For whom
Android developers working with AI agents and command-line workflows.
Problem
Agents generate outdated Android code and cannot test on real hardware.
Why now
Coding agents are mainstream in Android teams; Google is making its docs and device fleet agent-accessible.
Wedge
Official, CLI-first tools any agent can call, not tied to one IDE.
Money
Free; it keeps developers on Google tooling and its device-testing infrastructure.
Like
Firebase Test Lab · Gemini in Android Studio · Developer Knowledge MCP

LessonPlatform owners win the agent era by making their ground truth (docs, devices) machine-readable.

articleGoogle Developers BlogAndroid

Announcing ADK for Kotlin 1.0: Building Production-Ready AI Agents in Kotlin, Android, and Beyond

What
Version 1.0 of Google's Agent Development Kit for Kotlin, at feature parity with the Python and Java ADKs.
For whom
Kotlin and Android developers building agents on servers or inside apps.
Problem
Agent frameworks were Python-first, pushing Kotlin teams to switch languages or hand-roll HTTP calls.
Why now
In-app agentic workflows became an official Android pattern this week (Google's 'Build intelligent Android apps' series).
Wedge
Parity with Python, so Kotlin teams give up nothing by staying in their language.
Money
Free; it pulls agent workloads toward Gemini and Google Cloud.
Like
Koog (JetBrains) · LangChain4j · Spring AI

LessonReaching parity in a team's own language removes the main excuse not to adopt.

articleChris BanesAndroid

Haze 2.0

What
Haze 2.0 by Chris Banes: a Glass effect, a new effects API, performance work and experimental native Android backdrop support for Compose Multiplatform.
For whom
Compose and Compose Multiplatform developers building blurred or glass-style UI.
Problem
Background blur is hard to do well and fast across platforms in Compose.
Why now
Glass-style UI is in fashion and Android now offers a native backdrop API Haze can sit on.
Wedge
One API across Android, iOS and desktop.
Money
Open source; reputation and community.
Like
Compose progressive blur · Cloudy · SwiftUI materials

LessonLibrary authors stay relevant by adopting new platform APIs underneath while keeping their own API stable; his companion post on progressive blur shows how to explain that.

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

8
  • morluto/reacard

    Reverse engineer anything with agents, from app behavior down to native binaries.

    ★ 25.6k today

  • boykopovar/AnyPS5

    Automatic porting of PS5 executables to PC; tooling is collapsing the cost of binary porting.

    ★ 3,977 today

  • storytold/artcraft

    An open 'crafting engine' for artists who want control over AI image and video, not one-shot prompts.

    ★ 2,956 today

  • cathrynlavery/diagram-designcard

    Editorial diagram design for Claude Code, Codex, GitHub Copilot, Cursor, Factory Droid, and Pi. 44 diagram types. Self-contained HTML + SVG. No shadows. No Mermaid slop.

    ★ 1,083 today

  • bilawalsidhu/gods-eye-viewcard

    A spy satellite simulator in your browser, except the data is real. Live open source spatial intelligence on a photorealistic 3D globe.

    ★ 872 today

  • pytorch/pytorch

    PyTorch trending again; it is still the substrate under almost everything in today's HF columns.

    ★ 314 today

  • lyogavin/airllm

    Runs 70B models on a 4GB GPU by loading layers one at a time: speed traded for access on cheap hardware.

    ★ 541 today

  • chrislgarry/Apollo-11

    Apollo 11 guidance code resurfaces; famous source code is a reliable attention magnet.

    ★ 496 today

Product Hunt

50
  • Poke Finder

    A browser extension that turns browsing into a Pokédex hunt; engagement through collection, but IP-fragile.

    #1 on Product Hunt

  • kōdlo

    Sales workspace pitching one shared context across tools: the 'context layer' idea applied to sales.

    #2 on Product Hunt

  • Automations in Perplexity Computer

    Perplexity adds recurring background automations; assistants are becoming schedulers.

    #3 on Product Hunt

  • HireKey

    Job-search copilot whose angle is recruiter input rather than generic résumé rewriting.

    #4 on Product Hunt

  • Gboard Kurukuru Version

    Google Japan's annual joke keyboard; brand marketing more than product.

    #5 on Product Hunt

  • Foley

    Voice to-do list that asks before acting; 'ask first' as a trust feature for agents.

    #6 on Product Hunt

  • Ayven

    An agent that owns a team's meeting agenda; a painful but crowded wedge.

    #7 on Product Hunt

  • MePersona

    Replies to fans in a creator's voice; scales engagement but risks an authenticity backlash.

    #8 on Product Hunt

Show 42 more
  • Farolcard

    The macOS terminal that shows which coding agent needs you

    #9 on Product Hunt

  • Edgee Mod for Claude Codecard

    Use any model inside Claude Code, keep your Claude plan

    #10 on Product Hunt

  • NotchPit

    One of several notch utilities today; the MacBook notch has become a micro-platform.

    #11 on Product Hunt

  • KarmaBoard

    Split-flap message board for any TV; nostalgic design is the whole product.

    #12 on Product Hunt

  • DevTerminal

    Terminal with a project-aware Run button; IDE conveniences moving into the terminal where agents live.

    #13 on Product Hunt

  • Unboom

    Software room correction for Mac audio; DSP replacing acoustic panels.

    #14 on Product Hunt

  • Ghost Corecard

    A personal AI computer that runs entirely on-device

    #15 on Product Hunt

  • Kooler

    Cocktail finder from what's in your bar; the classic 'make it from what I have' utility.

    #16 on Product Hunt

  • Pickle Sensei

    AI swing coach for pickleball; vertical coaching riding a fast-growing sport.

    #17 on Product Hunt

  • SimpleSpread

    Native, open-source Mac spreadsheet; a bet that people want small native apps over web suites.

    #18 on Product Hunt

  • OpenScreen 2

    Screen recorder where an agent does the edits; editing becomes an instruction, not a timeline.

    #19 on Product Hunt

  • FocusWrite

    Local-first writing app; offline and privacy as positioning in a saturated category.

    #20 on Product Hunt

  • NewMinecraftServers

    Minecraft server directory with spoof detection; trust signals differentiate a listings market.

    #21 on Product Hunt

  • shippedit.me

    Launch platform selling SEO plus 'AEO' (optimising for AI answer engines), now a product category.

    #22 on Product Hunt

  • Tori

    Japanese listening practice generated from your own sentences; personal-content language learning.

    #23 on Product Hunt

  • Agibot

    Always-on Mac agent, free forever; price is the wedge and monetisation is unclear.

    #24 on Product Hunt

  • FloorGen

    Free floor plans from a sketch or a prompt; generative layout for non-designers.

    #25 on Product Hunt

  • Halo

    Notch app bundling notifications, dictation and agent status; the notch as an agent dashboard.

    #26 on Product Hunt

  • EdgeVolume

    Volume control from the trackpad edge; tiny utilities still launch well on Product Hunt.

    #27 on Product Hunt

  • Toolaby

    Two-line payments for Chrome extensions; picks and shovels for indie extension makers.

    #28 on Product Hunt

  • PixRater

    Fast photo culling and rating; a narrow professional workflow tool.

    #29 on Product Hunt

  • Naarchy

    A Dynamic-Island-style bar for Linux; the pattern spreading beyond Apple.

    #30 on Product Hunt

  • Cue

    Putting in AirPods triggers focus mode; context triggers as a product.

    #31 on Product Hunt

  • Skreno

    Record, edit and share video entirely in the browser; Loom's space with a no-install angle.

    #32 on Product Hunt

  • History Monk

    Browser history search, export and bulk delete; a power-user utility.

    #33 on Product Hunt

  • AdsNotch

    Rents your Mac's notch as ad space; novel, but likely to meet user and platform pushback.

    #34 on Product Hunt

  • Letra

    Copy any on-screen text on Mac via OCR; a long-tail utility that OS features keep threatening.

    #35 on Product Hunt

  • Museum of Models

    Archive of 371 models answering the same questions; a browsable comparison dataset.

    #36 on Product Hunt

  • Rank Kiwi

    Finds top-performing Instagram and YouTube content; creator analytics.

    #37 on Product Hunt

  • Firefox 157

    A major Firefox release launched on Product Hunt; incumbents use it for reach too.

    #38 on Product Hunt

  • Claude Dashboards & Motion

    Claude can generate live dashboards and animated explainers; chat outputs are becoming small apps.

    #39 on Product Hunt

  • Microsoft-Decision-1card

    Microsoft’s decision model for agents and workflows

    #40 on Product Hunt

  • ReSO AI

    Helps brands get recommended inside AI answers; the AEO category is filling quickly.

    #41 on Product Hunt

  • Onepin

    AI voiceover that pronounces names, prices and dates correctly; a precise pain point as the wedge.

    #42 on Product Hunt

  • Tuck

    Menu-bar organiser for macOS 27; OS releases reliably open utility gaps.

    #43 on Product Hunt

  • Isly

    Yet another notch app; this niche is crowding fast.

    #44 on Product Hunt

  • ej

    An 11MB local model for typed decisions in one pass; the decision-model trend at the tiny end.

    #45 on Product Hunt

  • put·here

    Capture now, organise later notes; the inbox pattern again.

    #46 on Product Hunt

  • Underplane

    A small browser game about spotting interceptions.

    #47 on Product Hunt

  • Buffer

    Clipboard history that also searches text inside screenshots; clipboard managers adding vision.

    #48 on Product Hunt

  • KernelAI

    Offline AI assistant for iPhone and Android; on-device as privacy positioning on mobile.

    #49 on Product Hunt

  • Fluence

    Listen to documents with on-phone speech synthesis; local TTS is now good enough to ship.

    #50 on Product Hunt

Show HN

5

HN front page

30
Show 22 more

YC Launches

20
Show 12 more

HF trending models

18
Show 10 more

HF trending spaces

14
Show 6 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

337 items37 sources7 seen in several sourcesall sources fetched

Source health

37Which sources were fetched today, and how many items each returned.
SourceSectionItemsStatus
GitHub TrendingMain8ok
Product HuntMain50ok
Show HNMain5ok
HN front pageMain30ok
YC LaunchesMain20ok
HF trending modelsMain20ok
HF trending spacesMain20ok
HF daily papersMain20ok
LobstersMain25ok
Dev.to topMain20ok
TechmemeMain15ok
AndroidX releasesAndroid · Official & releases33ok
Android Developers BlogAndroid · Official & releases4ok
Android Studio releasesAndroid · Official & releases4ok
Kotlin blogAndroid · Official & releases5ok
Kotlin releasesAndroid · Official & releases0ok
Compose Multiplatform releasesAndroid · Official & releases1ok
Android Developers YouTubeAndroid · Official & releases3ok
Google Developers BlogAndroid · Official & releases2ok
AndroidX merged API changesAndroid · AOSP / AndroidX insider3ok
Dove Letter (skydoves)Android · Experts5ok
skydoves on MediumAndroid · Experts0ok
Landscapist releasesAndroid · Experts2ok
Balloon releasesAndroid · Experts1ok
Jake WhartonAndroid · Experts0ok
Zac SweersAndroid · Experts0ok
Chris BanesAndroid · Experts2ok
Ryan HarterAndroid · Experts0ok
Talking KotlinAndroid · Experts0ok
Authority Insights (leaks & teardowns)Android · Leaks & ecosystem8ok
9to5Google · AndroidAndroid · Leaks & ecosystem6ok
Android AuthorityAndroid · Leaks & ecosystem6ok
Android FaithfulAndroid · Leaks & ecosystem1ok
Android WeeklyAndroid · Community15ok
GitHub Trending · KotlinAndroid · Community16ok
Lobsters · androidAndroid · Community2ok
HN · AndroidAndroid · Community2ok
1 of 5
#1Official

Android's command line tool can now drive real devices and guide coding agents

Why it matters Agents writing Android code can now test on real hardware and follow Google's own advice instead of outdated examples.

The Android Developers Blog announced two additions to the Android command line tool. Device Streaming lets you, or your coding agent, control real phones in Google's lab as if they were plugged in. Android skills are official instruction packs that ground agents in current guidance, including a new one for Wear OS with Compose Material 3.

Read the original · Android Developers Blog

1 of 4
#1Across sources

Microsoft ships a small model just for agent decisions

Why it matters Agents make many small calls per task, so a cheap, consistent judge can matter more than a smarter general model.

Microsoft released Microsoft Decision 1, a fast model that scores decisions inside AI agents and workflows, such as whether to approve, route or rank something. The first version is built on Qwen 3.5 with nine billion parameters, and Microsoft says it will move it onto its own and OpenAI's models. The same idea showed up today on Hacker News, at a Y Combinator startup and on Hugging Face.

Read the original · Product Hunt