tierlist.tech
中EN

09.25 Today's Picks

361 new projects found today, 203 rated.

Created 2026-09-23 – 2026-09-25Fetched 600Filtered 50Rated 31

S · Look today

S

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

A · Worth watching

A

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

A

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

A

wetype-ime-linux

yu1745C

Ports the WeChat input method's Android engine to Linux and integrates Fcitx5, filling the long-weak Chinese input on Linux; 156 stars in 1.2 days shows Chinese Linux users really lack this, with a chance to be preinstalled by distros and become a standard dependency. Risks are closed-source engine licensing and upstream version tracking, with almost no direct charging.

Linux port and Fcitx 5 integration for the WeType Android input engine

#Linux#Android#Reverse Engineering#Operating Systems

★156

A

binaural-voice

SaekisuiPython

Renders TTS voices with KU100 near-field HRIRs so they sound close to the ear and moving around the listener, with position labels, directly useful for spatial audio and voice demos; the scope is small and specific, more honest than big voice platforms, though listening validation is hard and it depends on specific HRIR data.

Make TTS voices speak right by your ear and walk around your head: KU100 near-field HRIR binaural rendering + position labels.

#Audio#Speech#3D

★44

A

ipadScreen

demirfiratSwift

It uses a jailbroken old iPad (tested on iPad 2 with iOS 6) as a Mac extended display, sending JPEG over USB/Wi-Fi with NEON decoding at about 35fps; the scope is small and the metrics concrete. It meets a real need for people hoarding old devices, but depends on jailbreak and old systems, and the experience gap versus Sidecar is unstated.

Use an old jailbroken iPad (iOS 6+, tested on iPad 2) as a real extended display for your Mac, over USB or Wi-Fi. Native iOS 6 app with libjpeg-turbo NEON, ~35 fps.

#iOS#macOS#Hardware#Performance

★14

A

morph-ui

AnmolSaini16TypeScript

Copyable animation components built on React 19 ViewTransition without pulling in an animation library, aimed at React developers wanting quick transition effects. The scope is small and clear, with code as documentation; the limitation is a shallow ecosystem and unknown component count.

Animated React components you copy and paste, built on the React's ViewTransition. No animation library, just React 19.3+ and a little CSS.

#Web Frontend#UI Design#Animation

★11

A

buried-injections

rudratoshsPython

Uses 629 AgentDojo injection samples to show regex and Prompt Guard 2 detect almost nothing in tool outputs, with reproducible scripts. Direct evidence for agent security practitioners, more useful than talk of defenses; but it covers only prompt injection and conclusions depend on the sample set.

🛡️ Regex catches 0%, Meta's Prompt Guard 2 catches 1% of 629 realistic AgentDojo injection attacks when they're buried in tool output. Reproducible benchmark.

#LLMs#Prompting#Benchmarks#App Security#AI Agents

★16

A

bitwright

valeria-orgRust

A Rust bit-vector expression simplification library for binary analysis and deobfuscation, filling an open-source gap beyond SiMBA. Its scope is restrained and its purpose stated in one line, so it can plug into reverse-engineering and vulnerability toolchains; the limitation is that it's brand new, with no docs or benchmarks.

Bit-vector expression simplifier for binary analysis and deobfuscation

#Reverse Engineering#Debugging#App Security

★20

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

BHas a spark43
  • B Seep-Reverse-Lab Turns multi-platform reverse engineering and CWE-602 client authorization bypass auditing into an agent-native workbench, letting models follow decompilation, protocol and auth logic, serving high-frequency mobile security and pentest grunt work; 88 stars shows security researchers willing to try. But IDA, Ghidra and Burp dominate, and big-vendor toolchains may absorb it. #Reverse Engineering#App Security#AI Agents#Exploits

    Agent-Native multi-platform reverse engineering and CWE-602 client-side authorization audit workbench.

    angusdevgoPython★88

  • B rooms Makes each project a room, switching the whole window layout for that project in one click, hitting the pain of repeatedly rearranging windows when running multiple macOS projects; 273 stars in 1 day shows real demand. But window management is a red ocean: Moom, Rectangle, free Raycast plugins and built-in Stage Manager, so independent buyout is hard to scale. #macOS#Productivity#UI Design#Automation

    Switch between projects on your Mac. Every project is a room: its windows, laid out, one shortcut away.

    saragordicSwift★273

  • B SuperMario3DWorld-VR A native OpenXR VR mod for Super Mario 3D World on Cemu, supporting stereo rendering, 6DOF, and first person. It's very specific for Wii U or VR players and technically substantial; the limitation is dependence on Cemu and the game itself, narrowing the audience. #Games#Game Dev#3D

    Native OpenXR VR mod for Super Mario 3D World (Wii U) on Cemu, with stereoscopic rendering, 6DOF head tracking, diorama and first-person modes.

    DestroyjevskiC++★30

  • B agents-dont-sleep A tray utility that prevents lid-close sleep only while agents like Claude Code or Codex are working, restoring afterward; the niche is tiny and the need real, spanning macOS/Windows/Linux, but it depends on recognizing each agent's process. #AI Agents#Productivity#macOS#Windows#Linux

    Close the lid, let your agents cook. A tiny tray app that keeps your laptop awake (lid shut, in your bag) only while Claude Code, Codex & friends are working, then lets it sleep. macOS · Windows · Lin…

    itsmbaqerRust★8

  • B voxel-musou A build-free browser voxel Musou game in Three.js, with Zhao Yun versus 300 soldiers. Both completeness and shareability are good, worth seeing for game devs and graphics fans, but it's a demo with limited gameplay depth. #Games#Game Dev#Web Frontend#3D

    Dynasty Warriors-style voxel action game in the browser: Zhao Yun vs 300 soldiers. Three.js, no build step.

    mike007jdJavaScript★65

  • B quiver-ae A billion-scale vector retrieval system for GPU and SSD, aimed at ANN infrastructure builders: spilling indexes that don't fit in VRAM to disk while keeping latency down, described without hype. The limitation is it's a pure research prototype with no install instructions or reproduction data, similar to FusionANNS and PipeANN. #GPU#Databases#Performance#Hardware

    A low-latency, high-throughput, and billion-scale GPU-SSD vector search system.

    ruc-systemCuda★22

  • B amfi-allow Lets ad-hoc signed binaries on macOS carry Apple private entitlements by cdhash without patching amfid, aimed at macOS developers doing signing or security research. The problem angle is very narrow and accurately described; the limitation is the high platform and permission barrier. #macOS#App Security#Reverse Engineering

    Let an ad-hoc signed binary carry Apple-private entitlements on macOS, by cdhash, without patching amfid

    Lakr233C★13

  • B pixel-office A Windows desktop app that puts multiple agents in a pixel office to collaborate, directly producing pptx/docx/xlsx/pdf, using your own key with approval before execution; the feature list is concrete, but the office-visualization wrapper has appeared many times this year, and reliability depends on how it handles failures and concurrency. #AI Agents#Windows#Productivity#Documents

    A Windows desktop app for visual AI agent teamwork in a pixel-art office. Create PowerPoint, Word, Excel and PDF files, switch AI models with your own API keys, and approve plans before agents act.

    ezmoney4scoutTypeScript★69

  • B ipc-edge A PLC data-collection and edge gateway for industrial sites, unifying multi-brand protocols with MQTT offline buffering and reporting, described specifically enough to judge who it suits; the limit is that such tools depend heavily on field protocol coverage, and the README lacks a supported PLC model list and deployment package. #IoT#DevOps#Networking#mqtt

    Cross-platform PLC communication and edge gateway tool — unified collection across multi-brand protocols, MQTT offline caching and reporting, collection and cloud upload in one program.

    agentthinkC#★11

  • B cookiebannerbench It quantifies the cost of consent banners across the same set of pages: appearance timing, size, viewport occlusion ratio, and layout shift, with metric definitions more concrete than saying experience is bad. The limitation is that the publisher is itself a CMP vendor, so sample selection and data need independent verification. #Benchmarks#Web Frontend#Privacy#Performance

    An open benchmark for what consent banners cost the pages they sit on — time to banner, payload, viewport blocked and page stability, measured on identical pages.

    cookieyesTypeScript★19

  • B digiemu An emulator running Elektron Digitakt mk1 firmware behind a clickable panel, with real-time audio and bundled .syx files; very appealing to niche hardware users and reverse-engineering/embedded enthusiasts, but limited to one model and unofficial. #Hardware#Audio#Reverse Engineering#Embedded#Windows

    Elektron Digitakt (mk1) emulator: runs the Digitakt's own firmware behind a clickable front panel, with live audio. Portable Windows app, bring your own .syx. Unofficial.

    irpinaPython★9

  • B ghost-editor An AI editing agent skill for talking-head short videos, featuring 7 styles, safe subtitles, and reverse-engineering from reference clips. The angle is specific and shareable, but it depends on HyperFrames and Claude Code, and actual output quality remains unverified. #AI#Video Editing#Agent Skills#Content Writing

    AI video editor for talking-head reels: 7 styles, face-safe captions, motion scenes, reverse-engineer any reference edit. A Claude Code / agent skill on HyperFrames.

    kurbaitaevJavaScript★41

  • B transgran-monstr Hashes phone numbers from CRM exports before sending them to large models while preserving relationships. Useful for teams needing AI customer-data analysis within compliance bounds, but it only covers Russian CRM scenarios and has few samples. #Privacy#Data Analysis#LLMs

    Anonymizes CRM exports before sending them to neural networks. Phone numbers become hashes, and joins survive.

    kurzemnekPython★8

  • B tev1 A decision model Together fine-tuned on Qwen3.5 4B with open weights. It's a small-model reference for people doing agent orchestration, and the institutional backing lends the stars some credibility, but there are no evals or usage instructions, so the payoff is unclear. #LLMs#AI Agents#ML Research

    Open-weight, Jev-inspired decision model finetuned on top of Qwen3.5 4B

    togethercomputerPython★82

  • B jevlint Writes semantic lint rules for code review issues that deterministic linters can't express, aimed at teams wanting to automate review. The angle is clear and it's not another formatter; the limitation is reliance on model judgment, with false positives and configuration cost still unseen. #Code Quality#AI#Testing

    Semantic lint rules for the code-review questions a deterministic linter can't express

    Ice-HazymoonTypeScript★13

  • B roblox-game-boilerplate An agent-friendly scaffold for Roblox projects: Rojo, Wally, typed networking packages, data stores, and AGENTS.md, more thought-out than most templates; but scaffolds themselves offer limited novelty, and value depends on whether these conventions keep up with Roblox and Rojo version changes. #Game Dev#Games#SDKs#AI Agents

    Roblox game boilerplate for AI agents: Rojo, Wally, typed Packet networking, ProfileStore data, strict Luau and an AGENTS.md

    frrazerLuau★16

  • B easyplot An agent skill for research data analysis that produces publication-grade charts in R or Python; the pain point is clear and it's more focused than generic plotting prompts, but it depends on an agent environment and an R/Python stack, and chart reproducibility isn't shown. #AI Agents#Data Analysis#Science

    AI Agent Skill for scientific data analysis and publication-ready figures in R and Python.

    RimaginationHTML★27

  • B glance-speedlab A local-camera inference latency test bench backed by measurement data, paired with the author's Glance. The methodology is honest and useful for on-device vision practitioners to compare against; but it's tightly bound to one product, so its value is limited without Glance. #Benchmarks#Performance#Multimodal#GPU

    A measurement-first lab for low-latency local camera inference with Glance.

    yoheinakajimaPython★35

  • B wow-ai It connects Claude, Codex, and Grok inside a World of Warcraft addon; since the addon sandbox cannot reach the network directly, there must be a bridge layer, which is worth a look. But the README is one sentence, with no communication method or ban risk, and 80+ stars look more like player curiosity than utility. #Games#LLMs#Plugins

    Addon to use AI (Claude, Codex and Grok) inside World of Warcraft (WoW)

    chelinho139JavaScript★83

  • B opencode-gpt-live Adds real-time voice calls to OpenCode, running GPT-Live via a ChatGPT subscription. Attractive for those wanting to dictate to a coding agent, but it relies on a subscription voice channel, so compliance and stability risks need checking first. #Speech#AI Agents#LLMs#Plugins

    Talk to OpenCode out loud: real-time GPT-Live voice calls on your ChatGPT subscription

    malhashemiTypeScript★20

  • B ollaya Locally pulls open decision models like Laya, decider, NLI, and GLiClass and serves them via a TypeSafe API, positioned as 'Ollama for decision models.' Promising for local classification/NLI serving, but the ecosystem and model quality remain unverified. #LLMs#APIs#Self-Hosted#ML Research

    Run open decision models locally: pull and serve Laya, decider, NLI and GLiClass behind a TypeSafe-compatible API. Ollama for decision models.

    ollaya-devRust★16

  • B muster-ww2 Procedurally generates WWII soldiers, tanks, artillery, and aircraft with three.js into a browsable 3D archive or sandbox, with a clear scenario for historical display and teaching and a restrained scope. The limitation is that it's a content project, with model accuracy and licensing unstated. #3D#Web Frontend#Learning

    Muster: a hands-on 3D archive of WWII military models. Browse, configure and arrange soldiers, tanks, artillery and aircraft on a diorama display. All models procedurally built in three.js + TypeScrip…

    Kenton-GMITypeScript★23

  • B Gaussian-Render-Capture A Blender add-on that renders models into COLMAP captures ready for training, feeding Gaussian Splatting pipelines. Useful for people making synthetic data, with a small clear scope; but it depends on the whole 3DGS toolchain, so the audience is narrow. #3D#Image Generation#GPU#blender

    Blender add-on: renders a model into a ready-to-train COLMAP capture for Gaussian Splatting (Postshot, LichtFeld Studio)

    virtualrepublicPython★9

  • B code-video An agent skill that researches a topic then generates a hand-drawn plus 8-bit style promo MP4, aimed at content or marketing people needing quick short videos. The workflow is concrete, but output quality and the required agent environment are unstated, so it reads like a demo. #AI Agents#Video#Content Writing

    Agent skill: research a topic and render a hand-drawn + 8-bit promo video (MP4)

    ChangroroJavaScript★27

  • B JEV-Paper-Radar Every morning a model reads arXiv and picks papers by interest, aimed at researchers tracking the frontier; cost estimates and fork-and-use are described concretely. The limitation is reliance on the Jev service, unverified recommendation quality, and plenty of similar tools already. #AI#ML Research#Science

    Let Jev read every new arXiv paper each morning and surface the few you should read. Plain-English interests, calibrated probabilities, ~$0.06/day, fork and go.

    Eliot5566Python★9

  • B claude-animation-skill Breaks hand-drawn 2D animation into detail settings, ant skeletons, watercolor brushes, and frame-by-frame reproduction for an agent to generate with canvas and ffmpeg; it's more specific than most skill packs, but output depends heavily on the model, and the repo lacks a complete film and failure samples. #Agent Skills#Animation#AI Agents#Video

    Hand-drawn 2D animation written as code: detail bible, a detailed ant rig, brush/watercolour pen, frame-matched recreation loop, synthesised sound. Node canvas + ffmpeg.

    buildwithhanifJavaScript★11

  • B MemBodied The official implementation of recurrent associative memory for vision-language-action models points in a direction useful for embodied AI; however, it is paper companion code without weights or training details, making external reproduction hard, so it suits researchers following along rather than direct use. #ML Research#Robotics#AI Agents

    Recurrent associative memory for vision-language-action models. Official implementation of MemBodied.

    declare-labPython★9

  • B win-plus-leave Locks the screen when a USB drive or YubiKey is pulled, closing the gap where Win+L gets forgotten, and can be deployed via Intune. The scenario is concrete and the PowerShell script easy to audit, but similar USB/Bluetooth lock tools already exist and the star count is very low. #Windows#Privacy#Productivity#yubikey

    🪟 + Leave: the Win+L you never forget. Locks Windows when your YubiKey or USB key leaves the laptop. Intune-ready.

    vdBurgITPowerShell★7

  • B sonic_for_a3 Agibot's SONIC adaptation for the full-size humanoid A3 targets researchers who own the robot or work on whole-body control. The official backing suggests the code may actually work, but a one-sentence description and missing startup instructions make it look like an internal engineering release. #Robotics#Hardware#SDKs

    SONIC adpation for new full size humanoid agibot A3！

    AgibotTechC++★30

  • B MKVDCU-Recomp An unofficial project recompiling Mortal Kombat vs. DC Universe to PC, aimed at game preservation or recompilation enthusiasts. The technical direction is specific, but with only one sentence it doesn't say compatibility or whether the original files are needed. #Game Porting#Games#Reverse Engineering

    An unofficial port of Mortal Kombat vs. DC Universe to PC (More platforms coming)

    KawaiiBungaC++★8

  • B SnafflePy A Python rewrite of the Windows credential-hunting tool Snaffler, filling a real gap for Linux-side internal pentesters; but the README is just "Snaffler in Python," with no rule coverage, run instructions, or differences from the original, and it can't even be installed today. #Network Security#App Security#Windows#snaffler

    Snaffler in Python

    S3cur3Th1sSh1tPython★53

  • B blender-claymation-skill A Claude skill that generates Blender stop-motion clay animation from one prompt, chaining storyboard, materials, and rendering; the direction is concrete, but it only gives prompts and a demo, with no render times, failure rates, or Blender version requirements, so usability must be tested yourself. #Agent Skills#3D#Animation#AI Agents

    Claude skill: stop-motion claymation films in Blender from a single prompt

    angrypenguinpngPython★8

  • B incremental It expresses object construction as a typed fold, where each step knows the exact shape and the final type proves completion, a rare convergent idea in type gymnastics; the audience is only heavy TS type users, similar small libraries abound, and the practical gain over a handwritten builder is unclear. #Web Frontend#SDKs#Code Quality

    A typed fold over object contributions: build an object piece by piece, know its exact shape at every step, and prove when it is complete.

    doeixdTypeScript★10

  • B blender-game-skills Claude Code agent skills for Blender game artists, turning concept art into riggable game assets and validating against reference images. The direction is specific and suits indie game teams, but it depends on the agent and a Blender environment. #Agent Skills#Game Dev#3D#blender

    Claude Code agent skills for game dev in Blender. Turn concept art into rigged, game-ready 3D assets, measured against your references.

    majidmanzarpourPython★13

  • B CVE-2026-84543 A technical disclosure and PoC for a macOS SMB kernel vulnerability, with a CVE number. Directly valuable to security researchers, but the number and impact need verification; ordinary developers should only read the disclosure, not run the PoC. #App Security#Exploits#macOS#Network Security

    Technical disclosure and PoC for CVE-2026-84543, a macOS SMB kernel vulnerability

    petermalonePython★22

  • B goutoujunshi-jev-chat A floating-window Mac tool that reads WeChat screens, analyzes conversation relationships, and drafts replies. The need is clear, but it screenshots private chats at no small privacy cost, doesn't explain model or data handling, and similar 'AI advisors' are already plentiful. #LLMs#Privacy#macOS#AI Agents

    Doghead Strategist Chat: Mac WeChat screen reading, relationship analysis and reply-draft floating window.

    shengjidaguai-chinaPython★23

  • B lifeline Remotely controls your local Cursor/CodeBuddy from a phone: view sessions, approve actions, dispatch tasks. The scenario is real, but solutions like Happy Coder already cover it, and the repo doesn't explain how the communication channel is secured. #AI Agents#Productivity#Plugins

    Remote-control the Cursor / CodeBuddy IDE on your machine — watch sessions, approve actions, and send tasks from your phone.

    zhangpaopao0609TypeScript★9

  • B container A TanStack project running a JS/TS sandbox in the browser, heading straight into StackBlitz and Sandpack territory; with one sentence and single-digit stars, it's unclear whether it targets doc embeds or IDEs, and no differences from WebContainers are shown, so just watch it for now. #Web Frontend#WebAssembly#Testing

    A browser-powered sandbox for JavaScript and TypeScript development.

    TanStackJavaScript★8

  • B openchatx-mcp Turns ChatGPT into a runtime that can control the local machine, call local tools, and forward to your own model via one MCP connection; the idea is on track, but MCP bridges are already plentiful, and it needs to show permission boundaries and tool registration, offering positioning only. #MCP#AI Agents#Automation#LLMs

    Turn ChatGPT into a local agent runtime. Operate your computer, use local tools, discover MCP servers, and delegate work to your own models through one MCP connection.

    XiaoPuOuOTypeScript★11

  • B shipvideo Enter a URL or a one-line request and it produces a product launch video, using an agent to write HTML and server-side rendering; this is a crowded AI-generated-materials space, and the README reads more like a landing page, with no rendering pipeline or controllability details, so output quality cannot be judged. #Video#AI Agents#Marketing#LLMs

    Paste a URL or a prompt, get a launch video. Opus 5.5 writes HTML, an OpenComputer serverless agent renders it.

    diggerhqTypeScript★24

  • B cve-2026-87902-wordpress-lfi-lab A reproduction environment for an unauthenticated LFI in WordPress, plus a scanner and PoC, with affected versions noted and an authorized-testing-only caveat, written properly. But such repos have very short lifespans, and value depends on whether the vulnerability is real and the PoC runs, so verify it yourself. #App Security#Exploits#Network Security#Debugging

    Reproduction lab + URL-list scanner + PoC for CVE-2026-87902 / GHSA-7hp8-65ch-5whp — WordPress get_page_template() unauthenticated LFI to conditional RCE (WP 4.7.0-7.1.1, fixed 7.1.2). Authorized/defe…

    dinosnPython★10

  • B zh-tech-writing An Agent Skill based on Ruan Yifeng's Chinese technical writing guidelines, emphasizing short sentences, plain language, and removing AI tone. Directly useful for Chinese technical writers, but it's essentially a prompt/spec wrapper, hardly a standalone tool. #Agent Skills#Content Writing#LLMs#Learning

    An Agent Skill for writing Chinese technical documentation, based on Ruan Yifeng's "Writing Style Guide for Chinese Technical Documentation": short sentences, plain language, no AI tone.

    leterLanguage unknown★38

  • B fathom A depth-style Alt+Tab for Omarchy that layers windows by recency and overviews workspaces. The interaction idea is fresh and scope restrained, but it only serves Omarchy users, limiting its generality. #UI Design#Linux#Productivity#omarchy

    A depth-based Alt+Tab for Omarchy: every window as deep as it is old, every workspace at a glance, any theme.

    mtolhuysQML★16

COrdinary65
  • C Relate A self-hosted private chief of staff that records what others said and what you promised from chat logs, targeting interpersonal memory rather than knowledge retrieval. The pain is real, but data access relies on gray chat-platform exports, privacy compliance is heavy, mem0 and other memory layers already own the developer market, and 97 stars won't convert to payment. #AI#Self-Hosted#Privacy#Knowledge Base

    A self-hosted AI chief of staff for your chats — remembers what they said and what you promised.

    georgedingJavaScript★97

  • C omarchy-meeting-recorder On Omarchy, records microphone and system audio on separate tracks, transcribes locally and generates speaker chapters, solving meeting-record privacy. But binding to one Arch distro narrows the audience, and local transcription already has cross-platform open-source rivals like Hyprnote and Meetily, so 101 stars can't support independent commercialization. #Linux#Speech#Privacy#Self-Hosted

    Record meetings on Omarchy: mic and computer audio as two tracks, transcribed on your own machine, with speakers, chapters and a player.

    jankeesvwRust★101

  • C short-video-generator-AI Uses an LLM to pick highlights, Whisper to transcribe and auto-editing to make vertical short videos, the most common open-source clone after Opus Clip's rise, differentiated only by being free. Whisper and ffmpeg have standardized the pipeline, so solo developers struggle to match mature SaaS on quality and asset compliance, leaving narrow commercial room. #Video#Video Editing#LLMs#Speech

    AI video processing pipeline for generating vertical shorts using LLMs, Whisper transcription, highlight detection and automated editing

    yukitoridoPython★188

  • C ClaudeAnimationBase Gives Claude a p5.js and p5.brush hand-drawn animation scaffold with 31 expressions and a prompt guide, essentially a model-capability template. It is a convenient starting point for short animation makers, but has no engine-level technology or paying audience, and prompt templates will be quickly replaced by better models and similar starter kits. #Animation#AI#LLMs#Web Frontend

    A starter kit for animating hand-painted cartoons with Claude: p5.js + p5.brush, the Clawd character, 31 acted emotions and a guide for the model

    JohnHeibelJavaScript★183

  • C bridgeclip Another open-source AI video clipping desktop app from creator BridgeMind, gaining early attention from channel traffic. But Opus Clip and a batch of similar open-source projects have made long-to-short standard, and it lacks a defensible model or distribution edge, likely staying a creator's own tool. #Video#Video Editing#AI#Editors

    Open-source AI video clipping desktop app by BridgeMind

    bridge-mindTypeScript★141

  • C MotionPosterior An implementation of motion-conditioned posterior estimation for small-object tracking, narrow but genuine research code; useful for multi-object tracking groups, but the repo has only a paper-style title, with no training config, weights, or reproduction commands, so outsiders can only treat it as a reference implementation. #ML Research#Images#Science

    MotionPosterior: Motion-Conditioned Posterior Estimation for Fast Small-Object Tracking

    ZSHYCPython★8

  • C dockdb-cnpj A Python tool for querying Brazil's CNPJ public company database with DuckDB, with a clear scenario and honest naming. It saves effort for those cleaning Brazilian company data, but it's region-specific and lacks performance notes for large data volumes. #Databases#Data Analysis#duckdb

    Consulta à base pública de CNPJ em DuckDB

    mateuspestanaPython★10

  • C painted-rickroll A one-minute animated oil painting drawn purely in code with WebGL2+WebAudio and composed music. The creativity is complete and implementation restrained; more an enjoyable creative-coding piece with low practical value, but worth ten minutes of study. #Web Frontend#Animation#Audio#Images

    Nocturne in Blue and Gold, the painted Rickroll: a one-minute moving oil painting with its own synthesised soundtrack, drawn entirely in code (WebGL2 + WebAudio).

    petergptHTML★12

  • C laya-rs A pure Rust runtime and HTTP service for small models like Laya System-1, CPU-first via ONNX Runtime. The direction is specific, but the target model is obscure and there are no performance numbers or usage examples; alternatives like candle and mistral.rs are more mature. #LLMs#CLI#Performance

    Pure Rust runtime + HTTP server for Laya System-1 models，Linux / CPU-first / multilingual / ONNX Runtime

    redwolf2019Rust★9

  • C Qualyx Claims a native port of Half-Life: Alyx to Quest headsets. If true that would be remarkable, but the repo doesn't say what's included or whether game assets are required, and its provenance and legality are opaque, so it's watch-only for now. #Game Porting#Games#Reverse Engineering#Hardware

    Native standalone Half-Life: Alyx for Meta Quest headsets.

    tinsarfalLanguage unknown★13

  • C disktree A disk-usage treemap written in Rust with GPUI, helping Omarchy users find and delete large files; the need is real but fully covered by dust, ncdu, Baobab and DaisyDisk, and it binds to one distro. A purely local small tool has no server or team-edition imagination and is hard to commercialize. #Linux#Performance#CLI

    A treemap for finding and removing what fills your disk, for Omarchy. Rust + GPUI.

    tobiRust★101

  • C NextForge An exploitation framework packaging common Next.js vulnerabilities, claiming coverage of RCE, SSRF, and middleware issues. It saves effort for authorized testers, but such scripts are readily available in security circles and there's no validation, so it's only for lab environments. #Exploits#App Security#Web Frontend#nextjs

    NextForge — unified Next.js exploitation framework. RCE · SSRF · middleware · authorized lab only.

    ynsmroztasPython★14

  • C zeroclaw-ui Adds a configuration UI on top of Zeroclaw, a companion interface to a single upstream project with no independent value; 201 stars in 0.6 days looks like spillover from upstream popularity. If upstream ships its own UI or changes the config protocol, this layer is immediately bypassed, and there is no chargeable angle. #UI Design#AI Agents#CLI

    UI for Zeroclaw setup

    CosmicchibattleTypeScript★201

  • C lucerna A desktop shell for Hyprland built with Quickshell and QML, aimed at Linux tiling window users. The target is clear, but with only a Portuguese description, 8 stars, and a day-old history, daily usability isn't proven. #Linux#UI Design#Productivity

    Shell de desktop próprio para o Hyprland, em Quickshell (QML)

    EduardoSA8006QML★8

  • C MathmoBench Yet another math reasoning benchmark, authored under a Cambridge math department name; math benchmarks have overflowed in the past two years, so if it's just a new question set, its value depends on contamination control and difficulty calibration, neither of which the README addresses. #Benchmarks#LLMs#ML Research#Science

    [Public preview] An advanced AI Benchmark for math, produced by mathmo at St John's College, Cambridge.

    Simreal-AIPython★33

  • C headless-dm Gives bots in an AzerothCore 3.3.5 private server origins, emotions, and news that shifts with rumors, a rare attempt to put LLMs into an existing game world; but the repo has only setting descriptions, no architecture or run steps, so how playable it is remains doubtful. #Games#AI Agents#LLMs#Self-Hosted

    Headless DM — an AzerothCore 3.3.5a realm where playerbots live as people in Azeroth: written pasts, lasting feelings, companies at war, and news that changes as it travels.

    bazolaPython★8

  • C wrayth A cyberpunk desktop shell for Hyprland built on Quickshell, a finished config from the ricing scene; the audience is clear but narrow, its value in visuals and plugin structure, but there are no dependency versions and no minimal setup steps beyond screenshots. #Linux#UI Design#Productivity

    A cyberpunk desktop shell for Hyprland, built on Quickshell.

    bowenbrideQML★11

  • C jev-chat-windows This tool reads WeChat windows with offline OCR, drafts replies with a model, and fills them back into the input box, avoiding hooks and database reads, which is useful reference for desktop-assistant builders. But the unofficial modified build may break at any time, and the privacy boundary for sending chat content to a model is unclear. #Windows#Automation#LLMs#Privacy

    Windows WeChat reply assistant (unofficial modified version): local offline OCR screen reading + model-drafted candidate replies + one-click fill into the WeChat input box, with optional auto-send. No…

    caizili999Python★8

  • C jevry An MIT desktop browser agent that completes web tasks for users, outputs citable research results and even plays games. The direction is already held by Browser Use, Comet and vendor computer-use features, and with heavy homogenization, 99 stars and no differentiated positioning mean no model edge, no distribution and very narrow charging room. #AI Agents#Browser Extensions#Automation

    Your browser. Ready to act. An MIT-licensed desktop browser agent for website tasks, cited research, and supported games.

    michaelswissaTypeScript★99

  • C parallio A multi-model desktop group chat plus Chrome extension letting several models converse in one session. The idea isn't new, similar aggregation clients abound, and the README doesn't clarify supported models or configuration; a day's star count isn't enough to judge. #LLMs#AI Agents#Browser Extensions

    Open-source multi-model desktop group chat and Chrome extension

    richard1208TypeScript★21

  • C Life-Compass-Obsidian-CN A Simplified Chinese localization of Compass aimed at Chinese Obsidian users. The author honestly labels it unofficial, but it's essentially template translation with no technical contribution; 68 stars in a day looks like community spread rather than project value. #Knowledge Base#Documents#Open-Source Alternatives#obsidian

    Unofficial Simplified Chinese localization of Compass; the original project was created by Daniel Agrici (@AgriciDaniel).

    sunkeke-aiJavaScript★68

  • C lichen A local, API-compatible alternative to Jev or System One for developers wanting to drop remote model dependencies; the direction is concrete, but a single sentence leaves model weights, performance, and compatibility range unstated, and 8 stars offer no usage evidence. #LLMs#APIs#Self-Hosted

    A local, API-compatible replacement for Jev, TypeSafe's System One model

    Mushroom-SystemsPython★8

  • C text-humanizer Uses multilingual LLM rewriting to make AI text sound human, essentially a prompt-engineering pipeline to evade AI detection; similar projects flood GitHub and are often banned by platforms. Free and open source with no paid entry, while detectors keep improving, it is an adversarial short-lived tool. #Content Writing#LLMs#Prompting

    A completely free open-sourced project designed to humanize AI-generated text through a multilingual LLM-powered rewriting pipeline.

    asokurasuPython★168

  • C webos-launcher A minimal launcher for LG webOS 6 TVs that requires no root. Useful for users wanting to escape TV home-screen ads and lag, with a clear scope, but it only covers older webOS versions, limiting its audience. #Self-Hosted#UI Design#Web Frontend#webos

    Minimalist home-screen launcher for LG webOS TVs (webOS 6 / Chrome 79). No root required.

    partridgeworksTypeScript★11

  • C openapp Aims to turn a single-user web app into a multi-user workspace, using versioned Adapters as a control plane. The problem is real, but the description stays conceptual with no examples and no target app, making it hard to judge usability today. #Self-Hosted#Web Frontend#DevOps

    Open-source control plane for turning existing single-user web apps into independent multi-user workspaces through versioned Adapters.

    seekskyworldTypeScript★13

  • C icloud-mail A self-hosted iCloud Hide My Email manager and code-retrieval API, derived from another project with a NOTICE file. It saves effort for anyone managing mailboxes in bulk, but mailbox pools plus auto code retrieval naturally sit close to account farming, with blurred boundaries. #Self-Hosted#Email#Automation#APIs

    Self-hosted iCloud Hide My Email management, email pool and API code-retrieval service; a derivative of mangobubu/icloud-api, with sources and license in NOTICE.md.

    t508708Go★8

  • C bestvaluemodel A daily-updated comparison page charting model intelligence indices against API prices to help people pick the best value model. The information is useful, but such rankings are plentiful and the implementation is just a static page, hardly a tool. #LLMs#Data Analysis#Benchmarks

    Which LLM is worth it: Artificial Analysis Intelligence Index vs API price, refreshed daily

    terrydsHTML★8

  • C octop-browser One of four octop repos released together by Tencent Cloud, controlling an agent browser directly via CDP; "more accurate and reliable" has no benchmarks or use cases, and tradeoffs versus Playwright aren't stated, so it's hard to tell a product from a department showcase. #AI Agents#Automation#SDKs

    Agent-first lightweight browser automation — direct CDP, no Playwright, more accurate and more reliable.

    TencentCloudPython★9

  • C octop-gateway Abstracts IM channels into a unified message interface for agents, a real need; but four repos shipped the same day with identical "unified abstraction" descriptions, showing no supported platforms or message mapping, so integration cost is unknown, and open-source bridges already exist. #AI Agents#APIs#Self-Hosted

    Multi-platform IM channel bridge with unified message abstraction for AI agents and bots

    TencentCloudPython★10

  • C small3d A pure C, zero-dependency software 3D renderer that draws models into a window. Good for anyone wanting to read a short rasterizer implementation, where the code is the documentation; but it has no practical use as a tool, and similar teaching projects are common. #3D#Learning#Performance

    software 3D object renderer using zero libraries in C

    teten-catC★11

  • C JevGuide Reads WeChat chats via the accessibility tree or screenshots so a model can judge relationships and draft replies, aimed at Android users; not modifying the client is a plus, but screen reading raises privacy concerns, and effectiveness and compliance risks are unstated. #AI#Android#Privacy

    Reading Between the Lines — a relationship-progress assistant for WeChat chats: screen reading (accessibility tree / screenshot vision) → Jev interpretation + progress score → chat model produces 3 ca…

    Nisaka520Kotlin★9

  • C octop-memory Cross-session memory for agents, not a new direction since mature projects already do it; the description stays at "helps you remember what matters," with no storage form, recall method, or conflict handling, making it the most marketing-like of the four same-batch repos, so skip it without an integration example. #AI Agents#RAG#Self-Hosted

    Help your agent remember what matters. Keep preferences, carry context across sessions, and bring your memory to the next agent.

    TencentCloudPython★9

  • C tidewater A 3D coastal town generated in one shot with Opus 5.5, essentially a model-capability demo rather than a reusable tool; 413 stars reflects curiosity about AI directly producing playable scenes. With no API, workflow or business path, such demos will quickly depreciate as models iterate. #3D#Web Frontend#Game Dev

    Coastal town built with Opus 5.5

    dgreenheckJavaScript★413

  • C kino-light An Android TV video and live-stream player built with Kotlin/Compose, with a clear target device. But such players are common, and the repo doesn't explain content sources, decoding capabilities, or how it differs from Kodi/Stremio, so novelty is thin. #Streaming#Android#Video

    Kino: reproductor de peliculas, series y TV en vivo para celular y TV (Android TV / Fire TV). Kotlin + Jetpack Compose.

    kinotvappKotlin★37

  • C 1point3acres-toolkit Automates 1point3acres check-ins and daily quizzes in a local Chrome, plus an MCP service for Claude Code. The implementation is transparent and bypasses third parties, but it's essentially an auto check-in script, many exist, and it skirts site rules. #MCP#Automation#Browser Extensions

    1point3acres daily check-in & quiz automation, local-only, plus an MCP server for AI assistants.

    vivian-labsPython★10

  • C agentrun A workflow DSL for Agentrun, aimed at developers orchestrating agent steps; the DSL is a worthwhile angle, but the repo is one sentence, showing no syntax, runtime, or differences from existing orchestration tools, and 29 stars prove little. #AI Agents#Automation#SDKs

    The Agentrun Workflow DSL

    Parcha-aiJavaScript★29

  • C expo-composer-hold-menu A React Native chat input component that long-presses to show a recent-photos picker, mimicking ChatGPT's interaction. A directly copyable part for Expo app developers, but the scope is tiny and the emoji-laden description limits its value to UI reference. #Web Frontend#UI Design#Plugins

    🚀 ChatGPT-inspired chat composer with a press-and-hold recent photo picker.

    rit3zhTypeScript★16

  • C Strata Claims to run a 125B MoE on 12-24GB VRAM plus 64GB RAM and serve an OpenAI- or Anthropic-compatible local API; if true that's useful, but the parameters and hardware demands are startling for one sentence, with no benchmarks or install details, so treat it with suspicion. #LLMs#GPU#Self-Hosted

    Qwen3.8-Flash-Next (125B MoE) on a 12-24 GB NVIDIA GPU + 64 GB RAM: one-click install for Windows / Linux. Strata inference engine, OpenAI/Anthropic API on localhost, optional image input.

    Niko1221C++★8

  • C ComfyUI-QwenImage-PhotoStyles A ComfyUI node for Qwen-Image photography styles, using a prompt rewriter to apply 17 camera languages. The description honestly says it's not a model loader and suits ComfyUI users, but it's essentially a style prompt pack with limited novelty. #Image Generation#Prompting#Plugins#comfyui

    ComfyUI node for Qwen-Image-2.1: 17 photographic styles (camera language + mood) via a PE-T2I prompt rewriter. Not a model loader.

    pottokao-dotcomPython★11

  • C genpark-customer-health-credit-burn-tracker-skill Tracks prepaid credit burn rate and computes runway and a health score for SaaS finance or customer success. The idea is clear but it's a single-point skill, 7 stars in 0.2 days shows no real usage, and it closely resembles Metronome. #AI Agents#Fintech#Data Analysis

    Real-time prepaid credit burn velocity tracker calculating runway days, health scores, and upsell triggers inspired by Metronome

    Alpha-ParkPython★7

  • C genpark-pricing-model-iteration-simulator-skill Builds a side-by-side revenue simulator for seat-based, usage-based, and hybrid pricing, aimed at SaaS pricing discussions. The problem is specific, but it's more an analysis template, and with only hours of history there's no evidence it beats existing spreadsheet models. #AI Agents#Fintech#Spreadsheets

    Side-by-side seat-based, usage-based, and hybrid pricing revenue simulator with elasticity scoring inspired by Metronome

    Alpha-ParkPython★7

  • C ex-nihilo A single-file HTML generative audiovisual work, more an openable digital art piece than a reusable library. It has reference value for creative coders, but with no install or integration entry point, the audience is very narrow. #Animation#Audio#Web Frontend

    EX NIHILO — generative audiovisual piece in a single HTML file

    ErwannRobinHTML★8

  • C geektls A TLS fingerprint spoofing client in Go, aiming to mimic browser JA3/JA4/HTTP2/3 fingerprints; potentially useful to scraper and anti-detection developers, but the README is one line, lacking usage and boundary notes, and mature alternatives like utls already exist. #Network Security#Privacy#APIs

    spoof client tls(ja3,ja4,http2.0,http3.0....more)

    geekbyterGo★10

  • C CopyToBaleBot A small bot syncing Telegram and Bale channels on Telegram Serverless. The goal is clear and the implementation likely lightweight, but the audience is limited to Iranian dual-platform operators and it depends on platform policies. #Automation#Social#telegram#bale

    Telegram ⇄ Bale channel sync bot on Telegram Serverless

    mohamadkhoshnavaJavaScript★11

  • C phpclaw-monorepo A general PHP AI agent engine bundling tools, memory, guards, and a ReAct loop, claiming compatibility with multiple CMS/frameworks. The coverage is too broad, lacking core implementation and examples, so it reads as a positioning manifesto rather than an evaluable tool. #AI Agents#LLMs#Plugins

    The universal AI agent engine for PHP. Tools, memory, guards, and a ReAct loop already built. Works with Laravel, Symfony, WordPress, Drupal, Magento, Joomla, OpenCart, PrestaShop. Any LLM provider. Z…

    phpclaw-phpPHP★9

  • C undertone A small text box that tells you how a message sounds before you send it, for people who care about tone; the idea is neat, but it depends on Jev, has only one sentence, and general writing assistants could easily cover it. #AI#Content Writing#Social

    A text box that tells you how your message sounds before you send it. Powered by Jev.

    Nuu-maanTypeScript★8

  • C jevclip It extracts highlight clips from subtitles and scripts with quoted summaries, overlapping existing podcast and video clipping tools; the core dependencies reuse terms seen across the same batch of repos, and neither the model nor the segmentation strategy is visible, so usability stays at demo level. #Video#Video Editing#LLMs#Speech

    Jev-powered video highlights and cited summaries from subtitles and scripts

    cclankPython★21

  • C zlaya A CPU inference engine written in Zig, focused on speed and small size, technically interesting to anyone wanting a non-C++ implementation; but it does not say which Laya model it means, which operators it supports, or what it benchmarks against, so today one cannot tell if it runs a real model. #LLMs#Performance#ML Research

    A fast and compact CPU-only inference engine for Laya in Zig.

    dip-protoZig★8

  • C qafiyah An Arabic poetry reference library/app with a clear title, potentially useful for Arabic literature and NLP data needs. But it's closer to a content resource, and the repo shows no retrieval, API, or data licensing details. #Learning#Content Writing#Data Analysis

    The Arabic Poetry Reference

    raaqimorgTypeScript★49

  • C qingming-bianjing One prompt has two models each build a 3D Bianjing web page, an interesting comparison demo, but it's an artifact, not a tool: no scoring criteria, no reproduction steps, and model versions don't match reality; fine as a curiosity, not as a benchmark. #LLMs#Web Frontend#Benchmarks#3D

    Same prompt, two cities — a real-time 3D Bianjing benchmark for LLMs.

    SilentFleetKKHTML★8

  • C funkmykali Installs hundreds of Kali tools needed for OSCP, PNPT, and other certifications with one command, convenient for exam takers; but it's essentially an install-script collection, such scripts are everywhere, "zero errors" has no CI backing, and tool-version conflicts with Kali sources are unaddressed. #Network Security#Linux#Automation

    Full Kali Linux tool installer for HTB CPTS, OffSec OSCP, TCM Security PNPT/PJPT, INE eJPT/eCPPT, and TryHackMe PT1/PT2. Covers all 28 HTB Academy Penetration Tester Path modules. System-wide, zero-er…

    arif-offsecShell★8

  • C creator-lab Scrapes Instagram Reel scripts and breaks down hooks for comparison, aimed at content operators rather than developers; the workflow depends on Apify and third-party model keys, and the description mixes in the same nouns seen across these projects, looking more like a demo than a reusable tool. #Marketing#Content Writing#LLMs#Video

    Research Instagram Reel scripts with Apify, Jev, and Fireworks or Groq. Bring your own keys, compare hooks, and record the patterns.

    artemnovitckiiJavaScript★28

  • C coa-vfog A graphics mod adding volumetric fog and light shafts to the CoA 3.3.5a D3D9 client; the scope is clear, but it serves only a specific old game client, with limited generality and reusability, offering little to those outside that community. #Games#Game Dev#3D

    Volumetric fog and light shafts for the CoA 3.3.5a D3D9 client

    jealous-soundC++★10

  • C jev-chat-jarvis-ios An iOS custom keyboard that, on long-pressing a message, suggests intent, risk, and candidate replies; the idea is interesting, but it requires self-compiling with Xcode, only reads the clipboard, and both privacy and usability remain unverified, with clear Jev ecosystem traces. #iOS#AI#Privacy#Productivity

    iPhone keyboard edition: one custom keyboard that works across all chat apps — long-press to copy the other person's message, and the keyboard shows intent, risk and candidate replies; tap to insert i…

    jev-chatSwift★10

  • C mpv-youtube-search A Lua script adding YouTube search to mpv, aimed at people watching videos with mpv. The function is immediately clear, but the repo has no description or usage, and 7 stars over two days shows no maintenance signs, making it a small script. #Video#Plugins#CLI

    LunedorLua★7

  • C vault-tag-system Select a Markdown vault in the browser, tag each note with Jev, with high-confidence auto-accept and rollback. Interesting for knowledge-base maintainers, but it depends on the unfamiliar Jev and lacks accuracy and local privacy details. #Knowledge Base#Productivity#AI#LLMs

    Pick a Markdown vault in the browser and use Jev to judge article by article which tags should be applied. High-confidence ones are adopted automatically, the rest go to manual review, and it's rollba…

    mukiwuTypeScript★7

  • C paperclip-guide A companion install guide for a NetworkChuck video on building an AI IT department. It's tutorial notes whose value is tied to the video and a specific Paperclip version, with nothing to review outside that context. #Learning#AI Agents#DevOps

    Companion guide for the NetworkChuck Paperclip video: install Paperclip and build an AI IT department. Every step tested on Paperclip 2026.916.1.

    theNetworkChuckTypeScript★31

  • C octop-harness A "production-grade agent runtime built on Harness Engineering philosophy" where the whole description is adjectives, with nothing on what state it manages, how it deploys, or how it differs from existing agent orchestration frameworks; vision-only repos like this deserve the lowest score. #AI Agents#SDKs#Self-Hosted

    A production-grade agent runtime engineered from the Harness Engineering philosophy.

    TencentCloudPython★9

  • C serv-s-quick-video-clip-editor-sqvce- A quick Python clip tool for splitting, trimming, and merging, aimed at memes and game clips. It's a thin ffmpeg wrapper stuffed with 'solves all your problems' sales copy, and with so many similar tools, no trade-offs or differences stand out. #Video Editing#Video#CLI

    [OPEN SOURCE] Cut, trim, merge video clips quickly with ease without launching a heavy program. Made for memes, gaming clips and similar basic use cases. Tired of not having a reliable program to do t…

    s3rvillePython★14

  • C awesome-qwen-image A list of Qwen-Image models, quantizations, prompts, and LoRAs. It's a handy index for people doing image generation, but it's a resource collection with nothing of its own, and its content lifespan follows model versions. #Image Generation#Prompting#Knowledge Base

    Qwen-Image 2.1. Checkpoints, quants, prompt engines, LoRAs, and tooling

    wildminderLanguage unknown★16

  • C Dreamerland2api Wraps Dreamerland's guest image generation in an OpenAI-compatible FastAPI, aimed at developers wanting to reuse existing OpenAI clients. But it relies on an unofficial guest interface, raising stability and compliance concerns, and its value is capped by the upstream. #AI#Image Generation#APIs

    OpenAI-compatible FastAPI gateway over Dreamerland guest image generation

    HaizhuAIPython★8

  • C AI-POWERED-DUAL-LAYER-CAMPUS-BUS-AUTHENTICATION-AND-AUTOMATIC-NUMBER-PLATE-RECOGNITION-SYSTEM Combines YOLOv8 license plate recognition with school brand verification into a school bus access system, a specific but narrow problem. The long title and course-project feel suggest homework, and missing deployment instructions make it hard to reuse directly. #AI#Images#App Security

    An AI-powered campus bus authentication system that combines Automatic Number Plate Recognition (ANPR) and institutional branding verification. YOLOv8 detects vehicles and plates, while PaddleOCR extr…

    AathiniSenthilkumarPython★8

  • C morphe-patches An Android app patch set containing only stable release artifacts, with no patch source or supported versions; such patch repos are common today, offer little assessable information, and may cross app licensing boundaries, so it's not recommended for now. #Android#Plugins#Reverse Engineering

    Morphe android app patches (for apps that I use) - stable releases

    bearinmindcatKotlin★27

  • C jev-flow It turns typed workflows into a standalone editing studio, but the repo has no verifiable definition of what the core terms mean or where workflow inputs and outputs live; low-code orchestration tools are already plentiful, and this one's positioning is still vague. #Automation#LLMs#SDKs

    Standalone open-source studio for typed Jev workflows

    daltonrpjJavaScript★7

  • C pega-omni A Rust-written OpenAI-compatible voice service aiming to self-host TTS/ASR behind a unified API. The direction is clear but competitors abound, and the README gives no model support, latency, or deployment details, so it's just another serving shell. #Speech#APIs#Self-Hosted

    OpenAI-compatible speech serving in Rust

    pegainfer-projectRust★8

DSkip86
  • D Interpark- The README openly says it is free and open source with no liability for profit use; it is really a ticket-grabbing automation script serving scalper-style seat demand. Its lifecycle ends with platform risk-control updates, it has no technical accumulation or legitimate business model, and brings legal and account-ban risks. #Automation#E-commerce#Scams

    Open source and free for everyone to use; the tutorial is in the readme, figure it out yourself. If you use it for profit and run into problems, I'm not responsible.

    michelle11032424-jpgPython★93

  • D system-design-notes-zh Translates English system-design interview notes into Chinese for Chinese job seekers; 160 stars shows the niche audience is real, but the output is translation of others' content with no technical increment, API or paid point. It only fits as a community knowledge base and is not an investable target. #Learning#Knowledge Base

    Chinese translation of System Design Interview notes (Chinese translation of liquidslr/system-design-notes)

    lukeTheNeuromancerLanguage unknown★160

  • D taohuayuan A 3D web scene of the Peach Blossom Spring tale, leaning toward cultural display and a frontend demo. It's a reference for three.js visual work, but not a tool or library, lacks deployment instructions, and is useless to ordinary developers. #3D#Web Frontend#Animation

    Peach Blossom Spring 3D web page.

    JimLiuJavaScript★19

  • D ZCodium A telemetry-stripped audit fork of an editor, a understandable motive, but promises like "commit-by-commit review" can't be verified; upstream is little known, and it's unclear whether it can update or install without telemetry, so at least a comparison table and build artifact checks are needed. #Privacy#Editors#Monitoring

    ZCodium, Independent audit fork of ZCode (ZCode → ZCodium). All monitoring and telemetry removed; upstream commits reviewed one by one.

    ZCodium-projectTypeScript★55

  • D reflex A Rust/CUDA native GGUF inference engine focused on cold starts and an agent 'System 1' decision loop. The vision is large, but the README reads like a positioning manifesto, lacking benchmarks, model support, and a usable entry point. #LLMs#AI#Performance#GPU

    A high-performance, GGUF-native Rust & CUDA inference engine optimized for cold-start latency and real-time 'System 1' agent decision loops.

    lateos-aiRust★11

  • D BonDriver-RTLSDR A BonDriver adapter connecting RTL-SDR to TVTest for ISDB-T single-segment TV. Very specific for Japanese TV reception hobbyists, but the audience is extremely narrow and it lacks channel scanning and driver compatibility notes. #Hardware#Video#Windows

    TVTest BonDriver adapter for RTL-SDR ISDB-T one-seg

    mouseosC++★9

  • D Bilibili_Spider Pure Python scraping of Bilibili info, supporting multi-account concurrency and anti-risk-control measures, aimed at data collectors. The feature description is concrete, but such scripts depend heavily on site changes, and the compliance boundary of bypassing risk controls is left to the user. #Social#Data Analysis#Automation

    Bilibili info scraping implemented in pure Python algorithms, supporting concurrent multi-account use and anti-risk-control.

    MARCHREADPython★8

  • D Bev A reproducible Persian-language evaluation using three-valued decision scoring and Jevfire-style reasoning, aimed at a very small research niche; the jargon is stacked and benchmarks and usage are missing, so outsiders can hardly tell what problem it solves. #ML Research#LLMs#Benchmarks

    Ternary decision scoring with Jevfire-style inference, typed APIs, and reproducible Persian evaluation.

    Reza2knPython★11

  • D 18-floors-of-mayhem A pixel multiplayer level demo assembled from Codex and command-line tools, worth a look as an "AI writes games" showcase; but it's a work, not a reusable project, with no level design or networking implementation notes, so neither players nor developers can take away a method. #Games#Game Dev#AI

    Pixel-art multiplayer descent and boss-battle game built with Codex and Combos CLI.

    andyhuo520JavaScript★16

  • D aide Claims to be an AI+IDE but only mentions DeepSeek support, with other models pushed to 27Q2, garbled syntax, and no screenshots or setup instructions. AI editors are already crowded, and this repo offers nothing to evaluate. #LLMs#Editors#AI Agents

    a ai+ide for deepseek herness（Other model will support in 27Q2）

    skyelan1999Go★21

  • D miliastra-beyond-simulator Described as a Lua-driven 2D sandbox but also tagged with DeepSeek and MCP, spanning game simulation to AI tools, making the core purpose hard to see. Target users and startup steps are unstated, so it reads more like a personal experiment. #MCP#Games#LLMs

    Thousand Stars Sandbox Simulator: a Lua-driven 2D external sandbox for the Wonderland, supporting DeepSeek Harness, Web and MCP.

    1475505JavaScript★11

  • D open-jev-laya A Laya multilingual browser demo built on Transformers.js, emphasizing pure front-end operation; it is more a technical demo than a tool solving a specific problem, lacking details on model size, language coverage, and usage boundaries. #AI#Web Frontend#WebAssembly

    Browser-only Laya multilingual demos powered by Transformers.js

    killkliPython★10

  • D jev-advisor A local decision-advice skill for Codex that claims to rank tools, estimate risk, and check completion evidence; the scope is too broad, with no concrete rules or examples, so it looks like prompt wrapping and its real benefit is hard to judge. #AI Agents#Agent Skills#Productivity

    A private, local decision-advice Skill for Codex that ranks tool choices, estimates risks, guides recovery, and checks completion evidence.

    Noe1120Python★7

  • D MuseFi A financial workspace for the Muse agent covering research, risk limits, order drafts, and a local broker, even memecoins and sports betting; the scope is too large and involves real trading, the description reads like a vision, and safety boundaries and runnable details are missing. #AI Agents#Fintech#Crypto

    MuseFi is an open-source financial workspace for Muse. One pack gives the agent a desk language for research, risk caps, ticket drafts, and a local broker that can fill a book — stocks, memecoins, Pol…

    PawlanDavidPython★11

  • D CCleaner-Pro-for-Win-macOS A piracy repository directly distributing CCleaner Professional installers with no original code; 90 stars reflects only demand for free software, not technical value. It will be DMCA-taken down anytime, has no commercialization path or dependency potential, and is entirely non-investable on compliance grounds. #Piracy#Crack#Windows#macOS

    CCleaner Professional (Windows & macOS).

    UnkillaLanguage unknown★90

  • D eikos Claims to be an open-source, calibrated financial decision model, but the repo is Python code with no weights and no evaluation protocol; "calibrated" and "for trading" are unverifiable assertions, and finance-plus-model repos are prone to marketing, so it scores low as is. #Fintech#LLMs#ML Research

    Open, calibrated, single-pass typed-decision models (4B & 27B) for finance and trading

    caiovicentinoPython★15

  • D comfy-workflow-landing It explains how to run someone else's ComfyUI workflow on your own machine, aimed at agent readers, with no scripts; it is essentially documentation, and the real pitfalls of workflow migration, dependencies and model versions, are not covered. #Prompting#Learning#AI

    Procedure for landing someone else's ComfyUI workflow on your own machine. A guide for AI agents — no scripts.

    dseditorLanguage unknown★10

  • D ai-banking-support-copilot A bank customer-service copilot demo emphasizing intent classification and simulated knowledge-base retrieval; the scenario is clear but it reads more like a portfolio demo, lacking real data integration and deployment notes, and not differing much from existing customer-service bots. #AI#LLMs#RAG#Fintech

    A conversational AI banking copilot designed to deflect level-1 support tickets by classifying customer intents and simulating knowledge base retrieval.

    initPMTypeScript★18

  • D nanoMuse A personal AI agent inspired by Meta Muse, emphasizing a name, face, background continuation, and confirmation before irreversible actions. The ambition is oversized; mobile comes first with web/desktop undecided, so it currently looks more like vision packaging. #AI Agents#Android#Productivity#LLMs

    🐾 nanoMuse: an open-source personal AI agent inspired by Meta Muse. One agent with a name and a face that does things instead of answering questions, keeps working while the app is closed, remembers y…

    nano-museKotlin★9

  • D alice_skill A "drop it in and it works" skill pack, with no word on which skills it contains, what agent runtime it depends on, or how to trigger them; such collections appear daily, lack verifiable single-point value, and leave you with nothing after reading. #Agent Skills#AI Agents#Productivity

    Alice integration skill pack · drop it in and use it · Drop-in skill suite for AI coding agents.

    alicewe1Python★12

  • D portfolio-template Yet another portfolio template, selling a scrolling contour background, bento homepage, and 3D carousel; the content is all placeholders, making it a visual demo indistinguishable from dozens of free templates, and the stars look more like template-site traffic. #Web Frontend#UI Design#Animation

    Portfolio template: profile rail, animated contour background, bento home, 3D page carousel, phone layout. All content is placeholder.

    brewed-opsCSS★18

  • D away-together-starter A so-called AI starter pack: a visual guide, recorded demos, and one prompt; it is more teaching material than tool, and the same batch of repos repeatedly uses identical terms, lacking anything independently verifiable. #LLMs#Learning#Prompting

    Free Jev-inspired AI starter: visual guide, recorded travel demo, reusable prompt and small training recipe.

    earlyaidoptersHTML★12

  • D learn-jev-end-to-end A hands-on AI agent course aimed at Jev, promising 13 use cases and a single OpenRouter key; it is a tutorial/content repository rather than a directly reusable tool, and with an unfamiliar ecosystem name, its value is worth watching rather than adopting now. #Learning#AI Agents#LLMs

    Learn Jev end to end: a free hands-on course. Build 13 AI agent use cases with a fast brain (Jev) and a slow brain (LLM). One OpenRouter key.

    harshithsunkuJupyter Notebook★8

  • D ai-kyc-aml-assistant An AI KYC/AML compliance dashboard whose description covers ID verification and sanctions-list screening; the direction is specific, but it is unclear whether it is a complete product or a UI demo, and the seriousness of compliance systems is hard to prove with a few days' repository. #AI#Fintech#Documents

    An AI-assisted compliance dashboard for accelerating customer onboarding, featuring automated document verification and PEP/sanctions risk screening workflows.

    initPMTypeScript★18

  • D vibecheck A new library from jlowin, described only as letting code make the simplest decisions plus two emoji; it is unclear whether it is a decision library, a testing tool, or a joke project, and with no usage or scope, its value cannot be judged for now. #Code Quality#Testing#Productivity

    ✨✅ The easiest decisions your code will ever make.

    jlowinPython★16

  • D line-animation Only a repo name and language, with a README that doesn't explain its purpose, so it's unclear whether it's an animation library or a demo. There's no runnable entry point for developers, and a day-old repo shows no usage signs; not worth opening unless clarified. #Animation#Images#UI Design

    0rekiPython★16

  • D luraph-deobfuscator A Luraph v14-15 deobfuscation tutorial aimed at people handling Roblox or Luau scripts. Such content ages fast and expires easily, and the repo is more a walkthrough than a maintainable tool, so practical usability is limited. #Reverse Engineering#Games#Learning

    a complete walkthrough to deobfuscate luraph v14-15: a comprehensive deobfuscator guide.

    4fundsagent-sourceLuau★8

  • D world-pose-fall-intelligence It claims a multimodal world model and fall detection demo, but the description stays at the concept level, with no data, model, deployment, or evaluation details; stacked AI buzzwords make it hard to tell what it can do today. #AI#Multimodal#Health#Monitoring

    SentinelFlow - multimodal world-model and fall-detection intelligence showcase

    hanfeny7Python★14

  • D limbus-zh-10ch A personal AI translation patch for Limbus Company Chapter 10, with a very narrow target audience; it is essentially a game text replacement pack, lacking quality assurance and install instructions, and touching copyright gray areas, so it is unsuitable as an open-source recommendation. #Games#AI#Content Writing

    Limbus Company Chapter 10 Chinese localization patch (personal AI localization, a gap-filler for the City Zero Association localization).

    haool871Shell★8

  • D Its.Yellow.Flowers An interactive webpage celebrating spring, where watering grows a sunflower and triggers confetti rain. The animation is fun, but it's a personal romantic piece or practice work with no reusable tool value and a very narrow audience. #Web Frontend#Animation#Games

    🌻 Una experiencia web interactiva y romántica creada para celebrar el Día de la Primavera. Acompaña el crecimiento de una planta en su maceta regándola con amor, revelando mensajes especiales hasta fl…

    ramirolacciJavaScript★8

  • D excel-codex-bridge The description only says anti-dumbing-down plugin, and the title points to an Excel and Codex bridge, but it's unclear what specific problem it solves or how to install it. It lacks usable information for both Excel users and agent developers. #AI#Spreadsheets#Plugins

    Anti-dumbing-down plugin.

    KaixxruaPython★15

  • D trashcan It records steps for installing local models on a trash-can Mac Pro, essentially personal environment notes; somewhat useful to owners of the same old machine, but with no general conclusions or performance data, readers take away no reusable method. #macOS#Hardware#AI

    Bootstrapping and optimizing trash can mac pro for local ai

    cjtrowbridgeShell★9

  • D ai-personal-finance-intelligence An AI personal finance dashboard that auto-categorizes transactions and gives spending insights; the space is crowded, the description is a feature list, and with no model, data source, or privacy handling details, it is hard to differentiate from existing budgeting apps. #AI#Fintech#Data Analysis

    An AI-powered personal financial management (PFM) dashboard that automatically categorizes transactions and surfaces actionable spending insights.

    initPMTypeScript★18

  • D cash-flow-forecasting An interactive cash-flow forecasting tool that simulates income and expense changes and shows historical liquidity; the features are clear but common, and if the README is just a UI demo its value is limited, so real calculations and import paths need to be seen first. #Fintech#Data Analysis#Spreadsheets

    An interactive cash-flow forecasting tool that empowers users to simulate income or expense changes and visualize historical liquidity trends.

    initPMTypeScript★18

  • D mountain-flood-uav-optimization The title points to mountain flood drone optimization, but the description only says open source, free, and continuously updated, with no model, data, or run instructions. 43 stars can't make up for the missing information, so usability can't be judged. #Science#Data Analysis#Robotics

    Open source and free, continuously iterated and updated!

    Akun-pythonHTML★43

  • D JevSDSQL Claims to be a "self-developing SQL database" with semantic operators, but it's unclear what the operators do at the storage layer or how consistency is guaranteed; the description repeats the same jargon seen across these projects, and after reading you don't know what queries it can run, while DuckDB plus LLM solutions are already mature. #Databases#LLMs#Data Analysis

    A self-developing SQL database with JEV based semantic operators and natural language queries.

    SheltercosmoPython★9

  • D storycast A TypeScript project that generates stories from a topic, described in one sentence with no model, input/output, or deployment details; with AI writing tools flooding the space, it lacks verifiable novelty and onboarding information. #AI#Content Writing#Web Frontend

    Type a topic. Get a story.

    ilkerzgTypeScript★19

  • D owasp-design-codigo_tdc-sp-2026-09 Companion materials for a Portuguese conference talk on avoiding vulnerabilities at the design stage. For those who didn't attend, it's just a slide archive with nothing runnable, a content repo rather than an evaluable open-source project. #App Security#Learning#Documents

    Conteúdos da apresentação "OWASP no Design de Software: evitando vulnerabilidades antes do deploy".

    renatogroffeC#★13

  • D Survey-AI-for-3D-modeling-Robotics An MIT CSAIL survey repo on 3D modeling, robotics, and frontier models, accompanying a paper. It has reading value for researchers but isn't an installable tool, so it's unsuitable as a today's open source project recommendation. #ML Research#3D#Robotics

    On the Opportunities and Risks of Frontier Models for 3D Modeling, Computational Design and Robotics (MIT CSAIL CDFG)

    MIT-CDFGHTML★11

  • D augmentor-agent Advertised as desktop and browser versions of Augmentor Agent, but the description mixes MIT with "resale requires authorization," leaving scope and licensing unclear; with no core capabilities or startup instructions, it can't be trusted. #AI Agents#Automation#Plugins

    Augmentor Agent Desktop and Browser. Public source under MIT with Augmentor Resale Restriction: free use and modification; resale requires permission.

    ManoloRemiddiPython★8

  • D auto-team The description only says rotating team, so it's unclear whether it's scheduling, team rotation, or game matchmaking. With no usage or scenarios, 26 stars doesn't explain what it does, so reusability can't be judged. #Collaboration#Automation#Productivity

    Rotating team.

    DaiXuYQJavaScript★26

  • D cpa-plugin-oai-basispoints The name suggests an OpenAI billing plugin for some CPA system, but the description is empty; it's unclear whether it's ad attribution, billing, or a proxy gateway. 53 stars in 0.3 days with no docs looks more like noise. #APIs#Fintech#Plugins

    JaxsonWangGo★53

  • D OnTheEdge A C tool that exports credentials from the Edge browser, aimed squarely at account passwords. It doesn't describe authorized testing scenarios, and the risk far outweighs any reusable value; it shouldn't be recommended to most developers. #Windows#Secrets#Reverse Engineering

    Edge Credential Dumper

    JssNGCC★8

  • D tiktok-reporting-bot An automation tool for mass-reporting TikTok accounts, with long keyword-stuffed sentences; the substance is abusing the platform's reporting mechanism, with no technical merit and compliance risk, so it scores low by rule and no one should touch it. #Scams#Automation#Social

    Automated compliance assistants are software-based systems engineered to organize, process, and streamline the submission of platform appeals or violation reports concerning TikTok profile, acc, and a…

    WorkerTrenchLanguage unknown★21

  • D prts-cursor A pictorial tutorial on installing a tool across systems, this is a content repository; the audience is narrowly limited by the title, and there are no scripts or configs to reuse directly, so per the rules it does not make the list. #Learning#Productivity#Editors

    Installation tutorials for each PRTS system.

    crowningggLanguage unknown★8

  • D SpringBoot2026 A Java sample repository named Spring Boot 2026, described only as for java spring boot project; it looks more like course or practice material, with no specific problem or onboarding path, and short-term stars are not a credible signal. #Learning#Web Frontend#APIs

    for java spring boot project

    hamitmizrakHTML★38

  • D xuanhuan-combat-director A prompt generator for xianxia/fantasy fight storyboards, piling on big terms like cinematic-grade and physics parameter libraries while adapting to multiple video models. It's a prompt template with exaggerated claims, making real output and maintainability hard to judge. #AI#Prompting#Content Writing#Image Generation

    Xuanhuan Combat Director — cinematic-grade action dynamics + VFX scale geometry + full move physics parameter library and character/scene presets, compatible with Seedance 2.5 / Kling / Jimeng. Produc…

    lixiaoxiao9888-createPython★7

  • D TypingPet Only a one-line Korean title, with no indication whether it's a typing pet or something else, and no visible tech stack. A repo whose purpose can't be determined isn't worth the time. #Games#Productivity#Learning

    Seungwoo's typing pet.

    swoonqxLanguage unknown★32

  • D EvoBot The repo has no description, so it's impossible to tell whether EvoBot is a bot, a game bot, or something else. 30 stars in just over a day lacks verifiable use cases, so it's not worth the time. #Automation#Games#Robotics

    ABC12345anouysPython★30

  • D CLM The repo is named CLM and written in Python, but has no description. It hit 807 stars in just over a day with no docs or entry point, suggesting star manipulation more than a real project; there's no verifiable information, so skip it. #ML Research#LLMs#Benchmarks

    Contrastive-LMPython★807

  • D Anthropic-Interview-Notes An interview-question bank that is really a content repo, not a tool: whether the questions are real and how broad the coverage is can't be verified, and its reusable value only holds for personal prep, so by rule it stays off today's list. #Learning#Jobs#Documents

    Anthropic Interview Questions and Answers

    SchuturePython★12

  • D chatgpt-prism2api The name suggests a Go service converting ChatGPT/Prism into an API, but the repo has no description, leaving licensing, stability, and compliance boundaries unclear. Such reverse-engineered wrappers carry high risk and aren't worth recommending. #APIs#LLMs#Reverse Engineering

    lvzhentaoGo★12

  • D xgoduck_hardware Only a repo name with no description; xgoduck looks like part of a bot or hardware project, but it's unclear whether it's CAD, a driver, or firmware. 71 stars in 0.6 days looks more like a batch release from the same organization. #Hardware#Robotics#Embedded

    LuwuDynamicsLanguage unknown★71

  • D OmniJev Empty description and empty README, yet over fifty stars in a day. There's no information about what it does, and the star count alone isn't evidence. #LLMs#AI Agents#Multimodal

    tinnel123666888Python★52

  • D Tiersense No description and no README, yet over thirty stars in a day, with no indication of what it does and no runnable entry point. #LLMs#Benchmarks#Data Analysis

    wangmingxuan666Language unknown★34

  • D sonovel--Android An Android novel app with an empty description, so it's impossible to tell whether it's a reader or a book-source scraper; both its technical form and purpose are unclear. #Android#Documents#Learning

    whykangKotlin★9

  • D Microsoft-Activation-Scripts The repo name directly borrows the well-known MAS activation script to bypass Windows/Office licensing; there's no assessable technical increment for readers, and it circumvents paid licensing, so it scores low by rule, with high name-clouting risk too. #Activation Bypass#Crack#Windows#Piracy

    Open-source Windows and Office activator featuring HWID, Ohook, TSforge, and Online KMS activation methods, along with advanced troubleshooting.

    Rolyata228Batchfile★33

  • D open-pi-viewer The repository has no description, and the name does not reveal whether it is Raspberry Pi video streaming, an image viewer, or something else; with no way to tell the target users or how to start it, just TypeScript code and few stars, it is not worth ten minutes. #Video#Images#Hardware

    gonzalez962TypeScript★9

  • D DIGITAL-FORENSICS An archive holding only a single project report, with no explanation of what it does and no visible code, so its purpose can't be judged and it isn't a reviewable open-source project. #App Security#Learning#Documents

    PROJECT REPORTS

    sivarama26Language unknown★7

  • D Self-Teacher-Context-Design The repo has no description or README, so it's unclear whether it's a method, dataset, or notes. It can't be evaluated and is best skipped. #LLMs#Prompting#Learning

    tiankanghuiPython★11

  • D dario-shabi Only a repo name and HTML, with an empty description; it's impossible to tell its purpose related to Dario or Shabi, and there's no runnable entry point. 18 stars in just over a day looks like casual starring, not worth opening. #Web Frontend#UI Design#Documents

    Darren-TerHTML★18

  • D Wardogs-Ultimate-Toolkit-2026 No description, and the name sounds like a game toolkit or cheat collection; its purpose and safety can't be confirmed, and the source is unclear. It has no value for serious developers, so skip it. #Game Cheats#Games#Windows

    DictatorRise34Python★11

  • D Gayatri-Tutor-ChemistryDemo A chemistry demo repo from an educational institution with an empty description; it's unclear whether it's course material, a demo app, or practice, so usability can't be assessed. It has no clear value for ordinary developers. #Learning#Science#Documents

    Gayatri-EducationLanguage unknown★12

  • D clawd-video No README content and no one-line description of what it does, just JavaScript code and a name; by rule a repo whose purpose can't be determined can't make the list, so wait until it explains what it does and how to run it. #Video#AI#Automation

    aadil6971JavaScript★8

  • D SoulStack Claims to give coding agents skills, agents, an immune system, and habits, complete with an official website domain. The terms are new but the concrete mechanisms and usage aren't explained, so it looks like an ecosystem marketing page rather than a hands-on open-source tool. #AI Agents#Agent Skills#Collaboration

    ProxySoul × Empryo: skills, agents, an immune system and habits for coding agents. soulstack.proxysoul.com

    proxysoulTypeScript★11

  • D browser-game-engine Empty description, so the engine form and target platforms are unclear. Browser engines are already plentiful, and this repo currently has nothing to evaluate. #Game Dev#Web Frontend#Games

    vinilanaJavaScript★7

  • D dataviz-skills A visualization skills repo with no description or README, so it's unclear whether it holds prompts, chart templates, or course material, and impossible to say who it's for. #Agent Skills#Data Analysis#Prompting

    z3ttLanguage unknown★7

  • D Aniimo-mod-1.9.0 No description, and the name sounds like a mod pack for some game or app; it's unclear whether it's a patch, assets, or an installer, and there's no compatibility info. 36 stars and 0.4 days of history aren't enough to prove value. #Games#Plugins#Windows

    LightServerChaosLanguage unknown★36

  • D imersao-arq-soft-ia-v1 Course project code from a training camp, meant to be followed along in class, with no reusable tool value for readers of this list; per the rules, coursework and practice projects score low. #Learning#AI#Web Frontend

    fullstackclubeducacaoTypeScript★8

  • D VOTTURS-ZCITY-FIXED The description is only REMOVED UNNECESSARY CODE, revealing nothing about whether it is a FiveM server fix, a game mod, or something else; with no feature description or install method, its credibility cannot be judged, so skip it. #Games#Game Dev#Plugins

    REMOVED UNNECESSARY CODE

    kazoo43Lua★13

  • D maschine-md-mm The repo has no description; the name only hints at Maschine/MD-MM. It doesn't say whether it's an audio tool, reverse engineering, or hardware adaptation, so its purpose can't be judged and it's skipped for insufficient information. #Audio#Hardware#Reverse Engineering

    mo0kidC++★15

  • D df-labexperiments A course record of digital forensics labs 1 through 10, belonging to lab assignments or notes rather than a reusable tool. It doesn't say who it's for or how to run it, offering no extra information to readers of the list. #Learning#App Security#Documents

    digital forensic laboratory experiments , from 1 to 10

    HimagirimolakalaLanguage unknown★14

  • D xgoduck_rl Another RL repo in the same series with an empty description; the training environment, algorithm, or task can't be determined, and there's no explanation of its link to the hardware repo. It offers no executable information for reinforcement learning readers. #ML Research#Robotics#AI

    LuwuDynamicsPython★13

  • D xgoduck_runtime_arduino An Arduino runtime repo with an empty description; it's unclear whether it's a library, example, or firmware, and there's no hardware model or wiring info. The series appears as a batch release, and this entry alone isn't worth opening. #Embedded#Hardware#Robotics

    LuwuDynamicsPython★7

  • D assignments A collection of assignments for a course, used by the author for teaching, with no transferable tool value for readers of this list; per the rules, coursework and practice projects score low. #Learning#ML Research#Science

    CS312 Deep Learning Alchemy student assignments.

    deep-learning-alchemyPython★8

  • D byeori The repository has no description, only Python code and a few stars; there is no way to know whether it is a library, script, or app, with neither target users nor an entry point, so based on current information it can only be skipped. #Productivity#Self-Hosted#Web Frontend

    joonan-labPython★11

  • D WebsiteStars No README description; the name hints at website star statistics, but functionality, data sources, and usage can't be confirmed. There's not enough information to judge, so it's skipped per the rules. #Web Frontend#Data Analysis#Productivity

    m2eatTypeScript★12

  • D skills The repo is named Paper Instruments Skills but has no description, so it's unclear whether it's agent skills, design resources, or something else. Only star counts exist with nothing to judge, so it's skipped outright. #Agent Skills#Productivity#UI Design

    Paper Instruments Skills

    paper-instrumentsLanguage unknown★25

  • D corrige-cp1 From the title it's just CP1 homework grading or correction material, a course or personal exercise not aimed at outside developers. There's no reusable tool or documentation, and 25 stars isn't evidence of value. #Learning#Documents#Knowledge Base

    Correção do CP1

    1TDSPA-26Language unknown★25

  • D CEB_Ardy The repository has no description, and neither the name nor the language reveals what it does; per the rule that undocumented purpose scores low, wait for the author to add a goal and minimal usage before evaluating. #Hardware#Embedded#Learning

    carlosedubarretoPython★10

  • D GoogleToThisCountry Both the repository name and description are a single identical phrase, revealing no specific problem it solves and no install or usage steps; per the rules, repos with undeterminable purpose do not make the list. #Networking#Linux#Privacy

    GoogleToThisCountry

    edmond1294Shell★10

  • D seo.web.cafe-extension Only the phrase 'Seo.Web.Cafe Extension' is given, without saying what the extension does or who it's for. Functionality and permissions can't be verified, so it's treated as an undocumented repo and skipped outright. #Browser Extensions#Marketing#Web Frontend

    Seo.Web.Cafe Extension

    qiayueJavaScript★9

  • D fuck0ai No description or README, and a title that's pure emotion, so it's unclear whether it's about bypassing payments, evading censorship, or something else. It can't be evaluated and shouldn't be recommended to readers. #AI#LLMs#Jailbreak

    zoowayssLanguage unknown★9

  • D Fortnite-Unlocker-Pro A Fortnite unlock or optimization suite claiming to deploy all cosmetics and lock offline profiles, making it a game cheat or unlock tool. It discusses no compliance risks and offers no technical explanation, so it shouldn't be recommended to developers. #Game Cheats#Games#Windows

    Automated Fortnite desktop unlock & optimization suite — full cosmetic deployment, FPS boost, offline profile locker. Windows 10/11.

    ImagevogalleryLanguage unknown★31

  • D facebook-reporting-bot A bot that automatically mass-submits Facebook reports or appeals; the description meanders, but the substance is abusing platform processes, with no legitimate use case, and it should not be recommended as an open-source tool. #Scams#Social#Automation

    Automated compliance assistants are software-based systems engineered to organize, process, and streamline the submission of platform appeals or violation reports concerning Facebook profile, page, an…

    NaturalCoreDevastateLanguage unknown★21

  • D Zona.Dev The repo has no description at all; the name only gives Zona.Dev, so it's unclear whether it's a website, tool, or exercise. With no README information to support judgment, it's skipped per the rules. #Web Frontend#Learning#Productivity

    ramirolacciTypeScript★8

Show HN 5

S

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

S

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

A

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

A

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

B

Show HN: Radix – Visual UI for agentic programming

Show HN

It turns one-off agent-generated artifacts into persistent, commentable, editable interactive widgets on local disk, trying to make the chat window no longer the only interface to an agent. The direction is right, but Claude artifacts, Canvas and notebooks already cover most scenarios, 15 points is low, and Pro revenue depends on replacing in-IDE experimentation.

Hey HN, I'm Jordan from Radix. Radix is a UI tool for programming agents. You prompt your agent to generate a workspace for a task you're working on and get an interactive widget that persists locally…

#AI Agents#UI Design#Web Frontend#Productivity

▲15

Hot New Repos

  1. mikehasa/golive-skill★874 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 GoLive account, backend or telemetry.
  2. Contrastive-LM/CLM★807
  3. riba2534/claude-opus-5-5-demo★778 claude-opus-5-5-demo
  4. yetone/magpie★611 Every agent's model. One place. Codex on DeepSeek, Claude Code on Kimi, from the menu bar.
  5. dgreenheck/tidewater★413 Coastal town built with Opus 5.5
  6. samyost1/3dicon★381 One prompt in, a looping animated 3D icon out — with real transparency. A Claude Code skill.
  7. 852wa/JIZURA★336 A browser app that automatically assembles text PVs from lyrics.
  8. saragordic/rooms★273 Switch between projects on your Mac. Every project is a room: its windows, laid out, one shortcut away.
  9. Aureliengmz/clearwater★265 Real-time, photoreal shallow water in a single HTML file. WebGL2, no libraries, no build step.
  10. amitshekhariitbhu/ai-engineering-course★258 AI Engineering Course - A free and complete AI Engineering Course to learn AI Engineering step by step - from Machine Learning, Neural Networks, and Transformers to LLMs, Fine-Tuning, RAG, AI Agents, LLM Inference, Evaluation, AI Safety, and AI System Design.

Star gain 1d

  1. NandhaKishorM/laya+2622★23003
  2. HarbourMasters/2ship2harkinian+2199★2242
  3. vectorize-io/hindsight+1663★26677
  4. eternity4719/HowToLiveBetter+1138★14149
  5. Finderchangchang/jev-chat-JARVIS+839★6061
  6. jev-chat/jev-chat-jarvis+838★6060
  7. pbakaus/impeccable+781★70687
  8. jaredpalmer/kev+770★6746
  9. browser-use/jev-ultrafast+762★19817
  10. cloudflare/security-audit-skill+752★22291