This media is not supported in your browser
VIEW IN TELEGRAM
π¦ makerviking/sindricad
SindriCAD
SindriCAD is the parametric designer that finally lets you add real surface textures to 3D printed parts. It is a free history-based modeler that lets you sketch shapes, extrude them into solids, and then carve detailed tactile patterns like knurling or honeycombs directly into the geometry. Unlike simple shading, these are actual raised bumps that a slicer will print as physical texture. It also handles multi-color printing for the Snapmaker U1, so you can assign different filament colors to specific faces and send the job straight to your printer. It is a fresh, open-source tool for anyone who wants parts that feel as good as they look.
π° https://news.ycombinator.com/item?id=49173432
π @hackernewsgithubprojects
SindriCAD
SindriCAD is the parametric designer that finally lets you add real surface textures to 3D printed parts. It is a free history-based modeler that lets you sketch shapes, extrude them into solids, and then carve detailed tactile patterns like knurling or honeycombs directly into the geometry. Unlike simple shading, these are actual raised bumps that a slicer will print as physical texture. It also handles multi-color printing for the Snapmaker U1, so you can assign different filament colors to specific faces and send the job straight to your printer. It is a fresh, open-source tool for anyone who wants parts that feel as good as they look.
π° https://news.ycombinator.com/item?id=49173432
π @hackernewsgithubprojects
This media is not supported in your browser
VIEW IN TELEGRAM
π¦ bomwiki/partmode
PartMode
Build precise, editable 3D parts and full assemblies directly in your browser with PartMode, a free open-source parametric CAD tool. It brings professional engineering features like sheet metal, welds, and STEP exports to the web without needing heavy desktop software. What makes it really cool is that it treats AI agents as first-class citizens, letting typed programs modify the exact same geometry youβre working on. You get editable part history, precise measurements, and structural drawings, all running locally in a local-first web interface that respects your privacy. Itβs a rare blend of serious engineering capability and developer-friendly extensibility that finally makes browser-based design feel like real software.
π° https://news.ycombinator.com/item?id=49219597
π @hackernewsgithubprojects
PartMode
Build precise, editable 3D parts and full assemblies directly in your browser with PartMode, a free open-source parametric CAD tool. It brings professional engineering features like sheet metal, welds, and STEP exports to the web without needing heavy desktop software. What makes it really cool is that it treats AI agents as first-class citizens, letting typed programs modify the exact same geometry youβre working on. You get editable part history, precise measurements, and structural drawings, all running locally in a local-first web interface that respects your privacy. Itβs a rare blend of serious engineering capability and developer-friendly extensibility that finally makes browser-based design feel like real software.
π° https://news.ycombinator.com/item?id=49219597
π @hackernewsgithubprojects
This media is not supported in your browser
VIEW IN TELEGRAM
π¦ domidoom/prometheus
Prometheus: Your Personal AI Assistant
Let you run a private, secure personal AI assistant that connects to your real life. Prometheus is a fork of a project called Warden that turns your computer into a smart helper. Instead of just chatting, it actually handles tasks. You can give it your calendar, email, and contacts, and it will sync everything automatically. It even lets you control your desktop and browse the web safely inside a secure container. The coolest part is how it keeps your data private. It runs locally, so your secrets never leave your machine. You can use it with major AI models or even run one yourself for free.
π @hackernewsgithubprojects
Prometheus: Your Personal AI Assistant
Let you run a private, secure personal AI assistant that connects to your real life. Prometheus is a fork of a project called Warden that turns your computer into a smart helper. Instead of just chatting, it actually handles tasks. You can give it your calendar, email, and contacts, and it will sync everything automatically. It even lets you control your desktop and browse the web safely inside a secure container. The coolest part is how it keeps your data private. It runs locally, so your secrets never leave your machine. You can use it with major AI models or even run one yourself for free.
π @hackernewsgithubprojects
Media is too big
VIEW IN TELEGRAM
π¦ domdoss/prometheus
Prometheus: Your AI Assistant with a Private Brain
Prometheus is the personal AI assistant that finally keeps your digital life private and organized. It is a secure, lightweight server that acts as a dedicated brain for your Claude or Ollama models. Instead of trusting cloud services with your data, this project lets you control your own smart assistant. It connects to your calendars, emails, and messages, handling reminders, scheduling, and notes right from a clean chat interface. The coolest part is that it runs locally or on your own server, meaning your sensitive information never leaves your control. It is like having a super-smart helper who respects your privacy and actually remembers what you need.
π @hackernewsgithubprojects
Prometheus: Your AI Assistant with a Private Brain
Prometheus is the personal AI assistant that finally keeps your digital life private and organized. It is a secure, lightweight server that acts as a dedicated brain for your Claude or Ollama models. Instead of trusting cloud services with your data, this project lets you control your own smart assistant. It connects to your calendars, emails, and messages, handling reminders, scheduling, and notes right from a clean chat interface. The coolest part is that it runs locally or on your own server, meaning your sensitive information never leaves your control. It is like having a super-smart helper who respects your privacy and actually remembers what you need.
π @hackernewsgithubprojects
This media is not supported in your browser
VIEW IN TELEGRAM
π¦ uber/submitqueue
SubmitQueue: Speculative Merges for Green Builds
SubmitQueue lets developers automatically merge code changes by testing them in parallel against predicted future states. Instead of waiting for one change to finish before starting the next, it speculatively rebuilds and validates multiple submissions simultaneously. When tests pass, the code lands instantly. If one change breaks the build, SubmitQueue isolates the problem and retries the others without human help. This keeps the main code branch stable even when many teams are pushing updates at once. It is a clever way to handle large-scale development without manual intervention.
π° https://news.ycombinator.com/item?id=49138084
π @hackernewsgithubprojects
SubmitQueue: Speculative Merges for Green Builds
SubmitQueue lets developers automatically merge code changes by testing them in parallel against predicted future states. Instead of waiting for one change to finish before starting the next, it speculatively rebuilds and validates multiple submissions simultaneously. When tests pass, the code lands instantly. If one change breaks the build, SubmitQueue isolates the problem and retries the others without human help. This keeps the main code branch stable even when many teams are pushing updates at once. It is a clever way to handle large-scale development without manual intervention.
π° https://news.ycombinator.com/item?id=49138084
π @hackernewsgithubprojects
Media is too big
VIEW IN TELEGRAM
π¦ domdoss/warden
Warden: Your Personal AI Assistant
Warden is the personal AI assistant that turns your computer into a smart, multi-tool helper. Itβs basically like having a private secretary who can read your emails, manage your calendar, and even browse the web for you. Built as a fork of Dockbox, it runs locally so your data stays yours. You can chat with it through a simple web interface or even over WhatsApp and Telegram. It uses models like Claude or Ollama to understand your requests and execute real tasks, like sending texts or scheduling meetings. Itβs a clever way to automate your daily digital life without handing your secrets to big tech companies.
π @hackernewsgithubprojects
Warden: Your Personal AI Assistant
Warden is the personal AI assistant that turns your computer into a smart, multi-tool helper. Itβs basically like having a private secretary who can read your emails, manage your calendar, and even browse the web for you. Built as a fork of Dockbox, it runs locally so your data stays yours. You can chat with it through a simple web interface or even over WhatsApp and Telegram. It uses models like Claude or Ollama to understand your requests and execute real tasks, like sending texts or scheduling meetings. Itβs a clever way to automate your daily digital life without handing your secrets to big tech companies.
π @hackernewsgithubprojects
This media is not supported in your browser
VIEW IN TELEGRAM
π¦ priyanshu-byte-coder/contextrot
Contextrot: Spot Coding Agent Decline
contextrot measures where your coding agent starts failing in your own sessions. It reads local transcripts to find the exact context size where your AI makes mistakes. You get a simple yes or no verdict on whether your setup needs changes. Everything stays private on your machine with no uploads. It is the fastest way to stop wasting tokens on degraded performance.
π° https://news.ycombinator.com/item?id=48794873
π @hackernewsgithubprojects
Contextrot: Spot Coding Agent Decline
contextrot measures where your coding agent starts failing in your own sessions. It reads local transcripts to find the exact context size where your AI makes mistakes. You get a simple yes or no verdict on whether your setup needs changes. Everything stays private on your machine with no uploads. It is the fastest way to stop wasting tokens on degraded performance.
π° https://news.ycombinator.com/item?id=48794873
π @hackernewsgithubprojects
This media is not supported in your browser
VIEW IN TELEGRAM
π¦ t8mars/comfyui-minimax-h3-audio-t8
ComfyUI MiniMax H3 Audio T8: Resumable Long Video
MiniMax H3 Audio T8 lets you generate long, synchronized video and audio by stitching together short segments instead of trying to render everything at once. Most AI video tools crash or run out of memory when you ask for anything longer than a few seconds. This project solves that by breaking the request into manageable chunks, saving each piece, and then linking them together into a smooth final file. You can even pause the process, review each segment, and only accept the ones that look good before moving to the next one.
π @hackernewsgithubprojects
ComfyUI MiniMax H3 Audio T8: Resumable Long Video
MiniMax H3 Audio T8 lets you generate long, synchronized video and audio by stitching together short segments instead of trying to render everything at once. Most AI video tools crash or run out of memory when you ask for anything longer than a few seconds. This project solves that by breaking the request into manageable chunks, saving each piece, and then linking them together into a smooth final file. You can even pause the process, review each segment, and only accept the ones that look good before moving to the next one.
π @hackernewsgithubprojects
This media is not supported in your browser
VIEW IN TELEGRAM
π¦ weberdak/transtractor-lib
Transtractor Lib
Transtractor is the universal bank statement parser that turns messy PDFs into clean data without using any artificial intelligence. Most tools rely on heavy AI models that are slow, expensive, and unpredictable. This project does the opposite by using a Rust-based engine that reads your bank PDF like a human would, looking for specific text positions and patterns to extract your transactions. You get a simple Python library that takes a PDF file and spits out a neat CSV or spreadsheet with all your dates, amounts, and descriptions perfectly organized. It even works in your browser through WebAssembly so you never have to upload sensitive financial files to a server.
π @hackernewsgithubprojects
Transtractor Lib
Transtractor is the universal bank statement parser that turns messy PDFs into clean data without using any artificial intelligence. Most tools rely on heavy AI models that are slow, expensive, and unpredictable. This project does the opposite by using a Rust-based engine that reads your bank PDF like a human would, looking for specific text positions and patterns to extract your transactions. You get a simple Python library that takes a PDF file and spits out a neat CSV or spreadsheet with all your dates, amounts, and descriptions perfectly organized. It even works in your browser through WebAssembly so you never have to upload sensitive financial files to a server.
π @hackernewsgithubprojects
This media is not supported in your browser
VIEW IN TELEGRAM
π¦ transtractor/transtractor-lib
Transtractor-lib
Transtractor-lib turns messy PDF bank statements into clean data without using any artificial intelligence. Built in Rust for speed, it uses strict rules to find transactions based on their exact position on the page rather than guessing from text. You can plug this into Python for analysis or run it in your web browser. It handles weird date formats and different bank layouts by just following the layout structure. It is fast, predictable, and completely free to use. Grab the code if you need reliable financial data extraction that actually makes sense.
π @hackernewsgithubprojects
Transtractor-lib
Transtractor-lib turns messy PDF bank statements into clean data without using any artificial intelligence. Built in Rust for speed, it uses strict rules to find transactions based on their exact position on the page rather than guessing from text. You can plug this into Python for analysis or run it in your web browser. It handles weird date formats and different bank layouts by just following the layout structure. It is fast, predictable, and completely free to use. Grab the code if you need reliable financial data extraction that actually makes sense.
π @hackernewsgithubprojects
This media is not supported in your browser
VIEW IN TELEGRAM
π¦ donchitos/claude-code-game-studios
Build Games with 49 AI Agents
Turn a single Claude Code session into a full game development studio. Instead of chatting with one general assistant that might write messy code, you get 49 specialized AI agents organized like a real studio. There are directors guarding the vision, department leads owning specific domains, and specialists doing the actual coding and design. It solves the problem of solo AI development by adding structure, quality checks, and a clear hierarchy. You still make every decision, but now you have a team that catches mistakes early and keeps everything organized from brainstorm to launch.
π @hackernewsgithubprojects
Build Games with 49 AI Agents
Turn a single Claude Code session into a full game development studio. Instead of chatting with one general assistant that might write messy code, you get 49 specialized AI agents organized like a real studio. There are directors guarding the vision, department leads owning specific domains, and specialists doing the actual coding and design. It solves the problem of solo AI development by adding structure, quality checks, and a clear hierarchy. You still make every decision, but now you have a team that catches mistakes early and keeps everything organized from brainstorm to launch.
π @hackernewsgithubprojects
This media is not supported in your browser
VIEW IN TELEGRAM
π¦ swaroop2004/proven-viral-content-system
Proven Viral Content System
Proven Viral Content System turns simple text prompts into ready-to-post videos for Instagram, TikTok, and YouTube without writing code. It is built for AI coding assistants like Cursor or Claude, using a chat-based workflow where you guide the process with simple commands. You start by setting up your brand, then the tool researches trending ideas and helps you brainstorm concepts tailored to your audience. The standout feature is its ability to generate actual MP4 video files by connecting to the Magic Hour API, handling everything from detailed prompt writing to downloading the final clip. You can even preview results without spending credits.
π @hackernewsgithubprojects
Proven Viral Content System
Proven Viral Content System turns simple text prompts into ready-to-post videos for Instagram, TikTok, and YouTube without writing code. It is built for AI coding assistants like Cursor or Claude, using a chat-based workflow where you guide the process with simple commands. You start by setting up your brand, then the tool researches trending ideas and helps you brainstorm concepts tailored to your audience. The standout feature is its ability to generate actual MP4 video files by connecting to the Magic Hour API, handling everything from detailed prompt writing to downloading the final clip. You can even preview results without spending credits.
π @hackernewsgithubprojects
This media is not supported in your browser
VIEW IN TELEGRAM
π¦ pradipta/wallfacer
Wallfacer: The AI Coding Session Manager
Wallfacer organizes the chaotic mess of AI coding sessions by creating a single searchable index of your work across Claude Code, Cursor, Kiro, and Codex. Instead of digging through scattered, unnamed transcript files on your hard drive, this tool reads them safely without touching the original data, then lets you name, tag, and group every conversation in a clean interface. You can instantly resume any past chat, rename old work, or search for specific prompts across all your agents in one place. It turns a week of scattered AI experiments into a structured, easy-to-find library of your actual progress.
π° https://news.ycombinator.com/item?id=49192219
π @hackernewsgithubprojects
Wallfacer: The AI Coding Session Manager
Wallfacer organizes the chaotic mess of AI coding sessions by creating a single searchable index of your work across Claude Code, Cursor, Kiro, and Codex. Instead of digging through scattered, unnamed transcript files on your hard drive, this tool reads them safely without touching the original data, then lets you name, tag, and group every conversation in a clean interface. You can instantly resume any past chat, rename old work, or search for specific prompts across all your agents in one place. It turns a week of scattered AI experiments into a structured, easy-to-find library of your actual progress.
π° https://news.ycombinator.com/item?id=49192219
π @hackernewsgithubprojects
This media is not supported in your browser
VIEW IN TELEGRAM
π¦ aymanbagabas/drift
Drift: A Git Diff Pager That Actually Wants to Be Looked At
Drift turns those boring command-line diffs into a vibrant, clickable terminal experience that makes code review actually enjoyable. Instead of scrolling through endless gray text, you get a full-screen interface with syntax highlighting, a resizable file sidebar, and word-level changes that pop with color. It even watches your repository in real-time, so you can see edits appear instantly while you work. It works as a drop-in replacement for your default git viewer, requiring zero setup to start using right away. Give your code reviews a visual upgrade today.
π° https://news.ycombinator.com/item?id=49206227
π @hackernewsgithubprojects
Drift: A Git Diff Pager That Actually Wants to Be Looked At
Drift turns those boring command-line diffs into a vibrant, clickable terminal experience that makes code review actually enjoyable. Instead of scrolling through endless gray text, you get a full-screen interface with syntax highlighting, a resizable file sidebar, and word-level changes that pop with color. It even watches your repository in real-time, so you can see edits appear instantly while you work. It works as a drop-in replacement for your default git viewer, requiring zero setup to start using right away. Give your code reviews a visual upgrade today.
π° https://news.ycombinator.com/item?id=49206227
π @hackernewsgithubprojects
This media is not supported in your browser
VIEW IN TELEGRAM
π¦ shawnpana/phone-harness
Phone Harness
You can let an artificial intelligence control your real iPhone directly from your Mac. This project bypasses jailbreaking and complex automation tools by using Apple's built-in iPhone Mirroring feature. It works like a digital puppeteer where a computer program takes screenshots through the mirroring window, reads the text on your screen using built-in vision tools, and then sends simulated taps and swipes to the phone. Essentially, it gives an AI agent eyes to see the screen and hands to click buttons, letting it open apps or type messages for you without needing special developer permissions or third-party drivers.
π @hackernewsgithubprojects
Phone Harness
You can let an artificial intelligence control your real iPhone directly from your Mac. This project bypasses jailbreaking and complex automation tools by using Apple's built-in iPhone Mirroring feature. It works like a digital puppeteer where a computer program takes screenshots through the mirroring window, reads the text on your screen using built-in vision tools, and then sends simulated taps and swipes to the phone. Essentially, it gives an AI agent eyes to see the screen and hands to click buttons, letting it open apps or type messages for you without needing special developer permissions or third-party drivers.
π @hackernewsgithubprojects
Media is too big
VIEW IN TELEGRAM
π¦ alvinreal/awesome-openclaw-tips
awesome-openclaw-tips
awesome-openclaw-tips is the practical guide that turns chaotic OpenClaw setups into reliable, cost-efficient workflows. Most people struggle with memory leaks and expensive token bills, but this repository solves that by providing tested, copy-paste tips for memory management, automation, and multi-agent systems. Instead of vague advice, you get exact config changes and implementation prompts you can run immediately. The coolest part is the heartbeat checklist system, which helps you schedule regular checks for email, calendar, and tasks without wasting compute. It basically teaches you how to keep your AI assistant working hard only when it needs to.
π @hackernewsgithubprojects
awesome-openclaw-tips
awesome-openclaw-tips is the practical guide that turns chaotic OpenClaw setups into reliable, cost-efficient workflows. Most people struggle with memory leaks and expensive token bills, but this repository solves that by providing tested, copy-paste tips for memory management, automation, and multi-agent systems. Instead of vague advice, you get exact config changes and implementation prompts you can run immediately. The coolest part is the heartbeat checklist system, which helps you schedule regular checks for email, calendar, and tasks without wasting compute. It basically teaches you how to keep your AI assistant working hard only when it needs to.
π @hackernewsgithubprojects
This media is not supported in your browser
VIEW IN TELEGRAM
π¦ datamllab/a2e
A2E Agent Auditing Engine
A2E is an open-source platform for auditing AI agents end to end. It lets you pair any benchmark dataset with any agent framework like LangChain or CrewAI and run experiments through a single command. The engine automatically captures every step, logging LLM calls and tool usage into detailed traces you can view in a local interface. Instead of just checking if the final answer is right, it scores the whole process, analyzing planning, efficiency, and tool selection. This gives developers full visibility into how their agents actually think, not just whether they passed the test.
π @hackernewsgithubprojects
A2E Agent Auditing Engine
A2E is an open-source platform for auditing AI agents end to end. It lets you pair any benchmark dataset with any agent framework like LangChain or CrewAI and run experiments through a single command. The engine automatically captures every step, logging LLM calls and tool usage into detailed traces you can view in a local interface. Instead of just checking if the final answer is right, it scores the whole process, analyzing planning, efficiency, and tool selection. This gives developers full visibility into how their agents actually think, not just whether they passed the test.
π @hackernewsgithubprojects
This media is not supported in your browser
VIEW IN TELEGRAM
π¦ tencent/yolo-master
YOLO-Master: Real-Time Object Detection
Tencent released a new research prototype called YOLO-Master that mixes two powerful ideas to make object detection faster. Instead of using one giant brain for every image, this system picks smaller, specialized experts only when they are needed. This dynamic approach saves computing power while keeping the speed high enough for real-time video. The code is based on official research presented at a major computer vision conference. It shows how splitting tasks between different neural network parts can improve performance without slowing things down. Developers can explore this open-source model to see how modern detection tools are evolving beyond standard single-model designs.
π @hackernewsgithubprojects
YOLO-Master: Real-Time Object Detection
Tencent released a new research prototype called YOLO-Master that mixes two powerful ideas to make object detection faster. Instead of using one giant brain for every image, this system picks smaller, specialized experts only when they are needed. This dynamic approach saves computing power while keeping the speed high enough for real-time video. The code is based on official research presented at a major computer vision conference. It shows how splitting tasks between different neural network parts can improve performance without slowing things down. Developers can explore this open-source model to see how modern detection tools are evolving beyond standard single-model designs.
π @hackernewsgithubprojects
This media is not supported in your browser
VIEW IN TELEGRAM
π¦ jtydhr88/comfyui-ultrashape1
ComfyUI UltraShape 1 Mesh Refinement
Take a rough 3D mesh and instantly sharpen it with ComfyUI UltraShape 1. This plugin adds image-guided refinement to your workflow, turning blurry or blocky geometry into high-quality, detailed models. You feed it a coarse mesh from another tool and a reference photo, and it uses deep learning to fill in the missing details and sharpen the edges. It works directly inside your existing ComfyUI setup, supporting common file formats and offering low-vram modes for smoother performance. This makes upgrading basic 3D assets to professional quality surprisingly simple and accessible for anyone building scenes or characters.
π @hackernewsgithubprojects
ComfyUI UltraShape 1 Mesh Refinement
Take a rough 3D mesh and instantly sharpen it with ComfyUI UltraShape 1. This plugin adds image-guided refinement to your workflow, turning blurry or blocky geometry into high-quality, detailed models. You feed it a coarse mesh from another tool and a reference photo, and it uses deep learning to fill in the missing details and sharpen the edges. It works directly inside your existing ComfyUI setup, supporting common file formats and offering low-vram modes for smoother performance. This makes upgrading basic 3D assets to professional quality surprisingly simple and accessible for anyone building scenes or characters.
π @hackernewsgithubprojects
This media is not supported in your browser
VIEW IN TELEGRAM
π¦ idollab/anytrack
AnyTrack: Tracking Objects with Any Data
Track moving objects in videos using any kind of data input you have. This project unifies visual object tracking by letting models adapt to different modalities like text, audio, or standard images. Instead of being locked into just seeing pixels, AnyTrack accepts various data types to identify and follow targets. It solves the problem of rigid tracking systems by offering flexibility across different sensory inputs. This makes it useful for complex scenes where visual cues alone might fail. The core idea is simple: give it any relevant data, and it finds the object. It is a fascinating step toward more adaptable computer vision tools that work with diverse information sources.
π @hackernewsgithubprojects
AnyTrack: Tracking Objects with Any Data
Track moving objects in videos using any kind of data input you have. This project unifies visual object tracking by letting models adapt to different modalities like text, audio, or standard images. Instead of being locked into just seeing pixels, AnyTrack accepts various data types to identify and follow targets. It solves the problem of rigid tracking systems by offering flexibility across different sensory inputs. This makes it useful for complex scenes where visual cues alone might fail. The core idea is simple: give it any relevant data, and it finds the object. It is a fascinating step toward more adaptable computer vision tools that work with diverse information sources.
π @hackernewsgithubprojects