#python #cactus #gemini #gemma #llm #on_device_ai

Needle 2 is a very small tool-calling AI model that can run on low memory, work offline, and return structured results like JSON instead of free text. It helps you build apps that call tools, extract data, and fine-tune models more safely and easily, even on devices with limited RAM.

https://github.com/cactus-compute/needle

#rust

Switchyard is a Rust proxy and library for LLM traffic that lets coding agents keep using their normal OpenAI or Anthropic API while requests are routed to different model backends and translated as needed. It can also split traffic for testing, measure latency and errors, and use custom routing rules, which helps you run or compare models more easily without changing your agent setup.

https://github.com/NVIDIA-NeMo/Switchyard

#swift #ai #dictation #ios #llama_cpp #macos #swift

FluidVoice is a free, open source macOS app that turns speech into text and can also control your Mac by voice. It works mostly on your Mac, so your voice and text stay private unless you choose cloud AI, and it supports fast dictation, live text preview, and optional smart formatting.

https://github.com/altic-dev/FluidVoice

#typescript #3d #ai_local #ai_tools #desktop_app #modly #open_source #self_hosted

Modly is a desktop app for Windows, Linux, and Apple Silicon Mac that turns a photo into a 3D model using open source AI on your own GPU. It supports extra model and process extensions, lets you run workflows, and includes a CLI for automation. This helps you make 3D meshes locally, with more privacy, faster control, and no need for cloud services.

https://github.com/lightningpixel/modly

#typescript #agent #agent_harness #agent_os #agentic #ai #ai_agent #ai_agents #artificial_intelligence #claude_code #codex #electron #holaboss #holaos #llm #mcp #memory #model_context_protocol #runtime #typescript #workspace

holaOS is a local-first desktop workspace that lets you run agents like Claude Code, Codex, or its built-in agent together with your tools, files, apps, and shared memory. It supports macOS, Windows, and Linux, includes built-in models or your own keys, and can help you get work done with real browser control, app automation, and ready-made files. This can save you time because one setup works across agents, keeps your context, and lets you produce useful results without switching systems.

https://github.com/holaboss-ai/holaOS

#other

This page is a guide list for using DeepSeek inside many AI tools and coding assistants, such as terminal apps, editor plugins, and desktop clients. It shows that you can set up DeepSeek in a few minutes, which helps you start coding or chatting with DeepSeek models faster and with less setup work.

https://github.com/deepseek-ai/awesome-deepseek-agent

#typescript #effect #framework #nodejs #plugin

Meta-Framework of Spatiotemporal Composability

https://github.com/cordiverse/cordis

#python #cli #consumer_gpu #dpo #fine_tuning #gguf #huggingface #llm #llmops #local_ai #local_llm #lora #low_vram #machine_learning #ollama #peft #python #pytorch #qlora #sft #transformers

Soup is a tool that helps you fine-tune and post-train large language models with one command, using one config file and little setup. It can work on a local GPU, even a 4 GB laptop GPU for an 8B model with layer streaming, so you can train without SSH or cloud hassle. This saves time, reduces setup pain, and makes model training easier to start and manage.

https://github.com/MakazhanAlpamys/Soup

#dart #android #exif #flutter #gallery #geotiff #gpx #metadata #metadata_explorer #motion_photos #mpf #slow_motion #svg #tiff #topojson #xmp

Aves is an Android gallery app that also shows photo details and metadata. It helps you browse, search, and map media files like JPEGs, MP4s, TIFFs, SVGs, panoramas, 360° videos, and more, so you can find and understand your pictures and videos faster.

https://github.com/deckerst/aves

#rust

ai-memory helps coding agents keep working without forgetting the plan, decisions, or failed tries between sessions. It saves notes in a simple markdown wiki, so the next agent can pick up the same task with a clear handoff. This saves time, avoids repeating setup, and makes it easier to continue work across different tools or days.

https://github.com/akitaonrails/ai-memory

#typescript

This is a starter project for a GenLayer football betting app. It includes a smart contract, fast local tests, full integration tests, contract checks, a Next.js frontend, and deploy tools. You can use it to build, test, and launch faster with fewer mistakes, because you can catch problems early and verify the app works before deployment.

https://github.com/genlayerlabs/genlayer-project-boilerplate

#plsql

This is an open-source radar project called AERIS-10. It is a low-cost 10.5 GHz phased array radar with two versions: one for 3 km and one for 20 km, using open hardware, FPGA processing, GPS/IMU support, and a Python control app. It helps you learn, build, test, or improve real radar systems with ready design files and code.

https://github.com/NawfalMotii79/PLFM_RADAR

#typescript #agents #claude_code #free #harness #harness_engineering #memory

This is a desktop app that turns coding CLIs into a team of agents that work together on your computer, with your own clone acting as the boss. It helps you by letting many agents share memory, send messages, and handle tasks while you can watch, guide, and approve important actions, so work can keep moving even when you are away.

https://github.com/chaitanyagiri/munder-difflin

#javascript #cloud #cloud_native #cncf #configuration_management #design_patterns #design_system #gatsbyjs #hacktoberfest #internship #kanvas #kubernetes #lfx #management_plane #meshery #orchestration #playground #visual_designer #visualization

Layer5 makes cloud software for managing apps and infrastructure, with tools like Meshery for Kubernetes, Kanvas for design and operations, Nighthawk for performance testing, a cloud native catalog, Academy, and shared patterns and components. The benefit is simpler teamwork, faster setup, better operations, and easier learning, so you can build and run cloud systems with less stress and more control.

https://github.com/layer5io/layer5

#rust

This text explains how to build, run, and test a local Amadeus node on Ubuntu with Docker or Podman, then start it as a system service with auto-updates and optional computor or validator startup. The main benefit is that you can quickly set up a working testnet, send transfers, deploy a contract, and run the node reliably on your machine.

https://github.com/amadeusprotocol/node

#mojo #ai #language #machine_learning #max #modular #mojo #programming_language

This repo has open-source tools for building and running AI with the Modular Platform, including MAX and Mojo. You can use it to start faster with model serving, AI code, examples, docs, and community help. It also shows where to contribute, report bugs, and check licenses, which helps you learn, build, and use the platform safely.

https://github.com/modular/modular

#python #agent #agent_security #ai_infra #ai_red_teaming #ai_security #llm #llm_evaluation #llm_jailbreak #llm_security #mcp_scan #openclaw_security #prompt_injection #prompt_security #scanner #security #security_tools #skill_scanner #skills_security #vulnerability

A.I.G (AI-Infra-Guard) is a Tencent Zhuque Lab tool for AI red teaming and security checks. It scans AI infrastructure, MCP servers, agent skills, and jailbreak risks, and it can run with Docker, a web UI, or an API. This helps you find weak points early, understand real security risks, and fix problems before they affect your AI systems.

https://github.com/Tencent/AI-Infra-Guard

#kotlin

Turn a Google Timeline file into a travel video on Android or iPhone, with exact date ranges, preview, and MP4 export. It works on-device, needs no Google sign-in, and keeps your Timeline file private. The benefit is fast, easy sharing of your trips as a video while keeping your location data under your control.

https://github.com/mahlernim/google-timeline-visualizer

#go

Agent Substrate is a system for running many agent-like apps at scale on Kubernetes. It can pause and restart them quickly, keep their memory and files, and pack many actors onto fewer machines, which saves cost and improves speed for you.

https://github.com/agent-substrate/substrate

#rust #dpi #gpui #hid #hidpp #local_first #logitech #logitech_mouse #logitech_options #mouse_remapping #mx_master #privacy #rust #smartshift

OpenLogi is a light, local-first, open-source replacement for Logitech Options+ that works on macOS, Linux, and Windows. It lets you remap buttons and keys, use profiles, control Logitech mice, keyboards, and webcams, and save settings in one plain TOML file. The benefit to you is more control, easier syncing between devices, and Linux support, with no account needed.

https://github.com/AprilNEA/OpenLogi