W39 This Week's Picks

2396 new projects found this period, 10 made the list.

Window 2026-09-21 – 2026-09-27Rated 1294

S · Look today

golive-skill

mikehasaTypeScript

Packages the dirtiest part of shipping agent-written code into an Agent Skill: domains, databases, email and payments all through the user's own accounts, with a zero-dependency CLI doing a reviewable detect→plan→apply flow. 874 stars in 1.4 days shows the ecosystem lacks this last mile, but Vercel and Cloudflare official Skills and built-in deployment directly cover it, making charging hard.

Take your agent-built product live: hosting, database, domain, email, payments — on your own accounts. Open-source Agent Skill + zero-dependency Node CLI: detect → plan → approve → apply → verify. No …

#AI Agents #Agent Skills #CLI #DevOps

★874

jev-chat-jarvis

jev-chatKotlin

Turns screen reading, intent judgment and candidate replies into a conversation copilot on the phone, solving the pain of not knowing how to reply in WeChat and Feishu, with no injection or packet interception reducing ban concerns; 3,723 stars in 1.5 days comes from strong话题 appeal. Monetization could be a subscription input-method-style entry point. Biggest risks: platform ToS and privacy compliance, and suspected star farming among sibling repos undermining credibility.

A conversation copilot installed on your phone: understand the other person in WeChat / QQ / X / Feishu, get candidate replies, and fill them into the input box with one tap — you decide whether to se…

#AI #LLMs #Productivity #Privacy

★3723

preflight

ghuntleyRust

A local proxy that scans LLM inputs and attachments for secrets before requests go out, addressing the frequent incident of employees pasting credentials into AI and leaking them. The author is influential and enterprise AI security budgets are clear, with a chance to become a DLP standard layer; competitors are vendors' built-in privacy filters and security gateways like Zscaler.

A local proxy that scans LLM requests and attachments for secrets before they reach the model.

#Secrets #Privacy #LLMs #CLI

★63

HaloBattery

HeyOkayPython

Shows battery levels for Logitech, Razer and other wireless mice, headsets and Bluetooth devices in one Windows tray, solving a very specific small annoyance. 170 stars in two days shows the pain is real; limits are Windows-only support and vendor protocol coverage you must read the README to learn.

Unified battery level indicators for wireless mice, headsets and Bluetooth devices in the Windows tray

#Windows #Hardware #Productivity

★168

A · Worth watching

Moon-FruityLoops26-FileParser

BridgeDruidCompressC++

Parses FL Studio's binary FLP project files in C++, enabling procedural generation, batch project edits, cross-DAW migration and AI composition training sets; 195 stars in 0.1 days shows the music toolchain long lacked such a low-level library, with potential as a format middleware via open core plus commercial SDK. Risks are format changes across versions and vendor legal pressure.

FLP binary parser for FL Studio project files

#Audio #Reverse Engineering #SDKs #Plugins

★195

cleanupper

SewCabinSpoutJavaScript

Solves the daily pain of macOS developers' disks being eaten by Xcode DerivedData, node_modules and Homebrew/pip caches, terminal-native, trash-first and zero telemetry, more trustworthy than CleanMyMac, and naturally distributable via Homebrew and social word of mouth. Commercialization could be a Pro rule library or team device cleanup, with the risk that Apple's built-in storage management and open-source peers like Mole already hold the space, and the anomalous 700+ stars in three days needs watching.

🧹 Free, open-source macOS disk cleanup CLI. Scan & safely remove caches, logs, Xcode DerivedData, npm/Homebrew/pip junk and stale node_modules from your terminal. Trash-first, zero telemetry the termi…

#macOS #CLI #Productivity

★723

ZCode

zai-orgTypeScript

Z.ai's officially open-sourced coding agent harness wraps its own coding model into a ready-to-use terminal agent, competing for the entry position of Claude Code/opencode. Over a thousand stars in 0.6 days comes from company channels and coding-agent heat. The commercial path is clear: funnel to its own API and subscriptions. Risks: severe homogenization, and harnesses easily swallowed by native model capabilities.

Z.ai's coding agent harness. Powerful, intelligent, extensible.

#AI Agents #LLMs #CLI

★1054

tpu-megakernels

InferactPython

Hand-fuses multiple TPU operators into a megakernel, targeting utilization loss from XLA compilation startup overhead and scheduling gaps, a scarce TPU performance skill. Inferact as an inference provider could package kernels as hosted inference or license them to cloud vendors, but Google's official Pallas and compiler keep absorbing such optimizations.

A collection of megakernels for TPU

#Performance #ML Research #Hardware

★95

agent-jev

malevrignsPython

A 0.6B System 1 decision model that maps states like diffs, traces and logs into calibrated probabilities in one forward pass, returning results in 50ms with zero decoding tokens and saving agent inference cost; 202 stars in 1.3 days, the asset in this family closest to a model proper, with potential as a standard agent dependency. Risks: confusing naming among peers and metrics lacking third-party validation.

AgentJev-0.6B - a fast 'System One' decision model for AI Agents: feed it any unstructured state (diffs, traces, logs) and structured questions, get calibrated probability distributions back in one ~5…

#AI #LLMs #AI Agents #ML Research

★202

arc-cua

shhivvPython

Provides a faster action-execution layer for computer-use agents, bypassing the slow loop of screenshot-then-reason at each step and directly targeting the latency bottleneck of GUI agents. 125 stars show CUA teams are looking for infrastructure, with potential to become the default action runtime; the risk is OpenAI and Anthropic's built-in toolchains absorbing this layer.

Superfast action layer for computer-use agents

#AI Agents #Automation #Multimodal

★125

That's today. The other tiers and the boards follow.

Show HN 10

Show HN: Whiteboard (YC W26) – An open-source IDE for thoughtful software design

Show HN

A CodeOSS-based desktop whiteboard where agents link architecture diagrams, sequence diagrams and code locations bidirectionally via SDK and Rust semantic diff folds noise, already used by Salesforce and Modal, hitting the cognitive debt of agentic coding. YC W26, MIT open source and a clear paid hosted path, with the risk of competing head-on with Cursor and Claude Code's built-in review.

Hello! We’re Sid, Alex, Ketan, and Milan. We’re building Whiteboard ( https://whiteboard.dev.fast/), an open-source desktop app where humans and agents can architect software together in a common work…

#AI Agents #Editors #Collaboration #Code Quality

▲166

Show HN: Foremerge – Catch intent conflicts between parallel coding agents

Show HN

Builds on git with intent and scope declarations to predict conflicts when multiple agents modify code in parallel, moving semantic conflicts from the PR stage to before writing code, with built-in MCP and acceptance gates. Multi-agent parallel work in one repo is a real pain point, and 451 stars plus 33 points shows engineering-circle resonance, with a chance to become a collaboration protocol standard; the risk is git vendors or Cursor building coordination in.

At, GPTree, we run several coding agents across our team on one repo using parallel worktrees. Apart from wasted time reviewing and fixing conflicts at PR time, the failures that hurt the most are whe…

#AI Agents #MCP #Code Quality

▲33

Show HN: An open-source manufacturing ERP/MES/QMS

Show HN

An open-source self-hosted manufacturing ERP+MES+QMS suite filling the gap for small factories bound by expensive closed SAP and Oracle systems; open-core and per-line licensing is a mature industrial-software path, and 51 points shows real demand. Risks are heavy implementation, consultant-dependent delivery, and thin compliance modules being squeezed by existing MES vendors.

#Self-Hosted #DevOps #Data Analysis

▲51

Show HN: Reladraw – A diagram language where you decide where to place things

Show HN

Fills the gap between Mermaid/Graphviz's lack of layout control and Draw.io's inefficient manual dragging: textual but position-controllable, and designed for agent editing. 232 stars in two days plus 151 HN points is the strongest signal in this batch, and the playground and Claude skill show real adoption. The business path is a design tool and standard entry point for agent drawing; the risk is Mermaid ecosystem inertia and Visio-like vendors following.

I love making diagrams to help understand, plan, etc. However, the options are (A) auto-placement languages like Mermaid or Graphviz (which don't let me decide how the diagram looks), or (B) software …

#UI Design #AI Agents #Agent Skills #Productivity

▲151

Show HN: Drop – A rootless Linux sandbox with gVisor support

Show HN

A rootless Linux sandbox using namespaces plus gVisor to give locally run untrusted dependencies and third-party programs a disposable isolated home, solving the pain of one package install compromising the whole machine; 153 points on HN shows strong developer resonance, turning around after two unnoticed launches thanks to gVisor support. Could serve as the execution base for AI-generated code. Risks: firejail/bubblewrap and containers already occupy the space, and it's Linux-only.

I created Drop because I always felt uneasy installing and running third-party programs using my main user account. A single compromised dependency means a full compromise of the system. What is even …

#Linux #App Security #DevOps #Privacy

▲153

Show HN: PlaceCall (YC W26) – agentic API to call businesses and get things done

Show HN

An API that gives agents phone calls: natural-language tasks, automatic IVR navigation and waiting on hold, returning structured results, recordings and transcripts to fix stale store info. YC W26, per-call pricing and 250 free tries are a standard API model; risks are voice infrastructure like Twilio moving down-stack and compliance around placing calls on users' behalf.

Hi HN, this is Yarik and Vlad from VOYGR - we are building the tools for agents and apps to engage with local businesses. It all started with our own pain point at VOYGR: calling businesses to verify …

#AI Agents #APIs #Speech #Automation

▲15

Show HN: I built a post-mortem debugger for native Windows x64/x86 crashes

Show HN

Post-mortem debugging for Windows x64/x86 crashes, moving WinDbg's analysis and cross-correlation into a modern GUI and using MCP so agents reason over structured data with fewer hallucinations—hitting a long-standing gap in native Windows troubleshooting. 20 points on HN and beta licenses already sold show willingness to pay. Competitors are WinDbg, Visual Studio and TTD; the risk is a closed-source small team struggling to shake the Microsoft ecosystem.

Hello HN! I've spent years debugging Windows crashes with tools that were either friendly but limited (e.g. Visual Studio) or powerful but archaic (e.g. WinDbg). I developed patterns and methods for u…

#Debugging #Reverse Engineering #Windows #MCP

▲20

Show HN: InstinctFlash – High-Performance Serving Runtime for Robotics Models

Show HN

An inference acceleration runtime for robot VLA and world models that cuts diffusion sampling from 25/50 steps to 2/4, directly addressing control-loop latency, a deployment bottleneck; backed by Jetson Thor measurements and Samsung/Siemens usage, though HN gave 23 points and GitHub only 42 stars, so momentum is just starting. Could become the vLLM of robotics. Risks: AGPL and pressure from NVIDIA's Isaac/TensorRT ecosystem.

Hey HN, Guanming here, cofounder of General Instinct. We just released InstinctFlash, a high-performance serving framework for robotics models. It’s licensed under AGPL-3.0. On Jetson Thor, we see spe…

#Robotics #GPU #Performance #ML Research

▲23

Show HN: AI·rete·RAG – a Rete rule engine decides, RAG explains why

Show HN

Uses a Rete rule engine to issue verdicts while the LLM only explains with cited policies, solving decisions that must be auditable like credit risk and clinical use, with Verdict reproducibility as a real selling point; 34 points on HN and an MCP client show early developer experimentation. The monetization path is clear (compliance audit SaaS). Risks: a closed-source engine squeezed by Drools/IBM ODM and built-in model guardrails.

Hi HN, I built ai·rete·rag because I kept seeing teams put an LLM in charge of decisions that need to be auditable (lending, fraud, clinical triage), then bolt on "guardrails" after the fact. It runs …

#RAG #MCP #AI

▲34

Show HN: Canary (YC) – Independent verification for AI code

Show HN

Independent verification for AI-generated code: takes a changeset and intent, then uses a swarm of agents in a remote sandbox to verify runtime behavior and security invariants, filling the gap between static review and unit tests. YC backing plus a per-repo CLI is a clear enterprise path. The risk is huge: Greptile, CodeRabbit and Qodo already hold the track, and only 5 HN points shows the narrative did not land.

Hey HN, we are Aakash and Viswesh and we are building Canary ( https://www.runcanary.ai/) - independent verification for AI code. Claude/Codex calls Canary with the changesets, intended behaviour and …

#AI #AI Agents #Testing #Code Quality

▲5

S · Look today

A · Worth watching

Show HN 10

Star gain 7d