Best Local AI AI skills

Llama.cpp enables fast and efficient inference of large language models in C/C++ applications, allowing developers to integrate AI capabilities into their projects

Stirling-PDF is a TypeScript-based application that allows users to edit PDFs on any device, anywhere, making it a versatile tool for managing and modifying PDF documents.

Run large AI models on your own PC: Qwen3.8-Flash-Next on Strata's engine, and GLM 5.3 Flash, DeepSeek V4 Flash, LFM2.5, Qwen3.6, Ornith 1.5, Gemma 4, Tiel Coder and MiniCPM-V on a second engine written for PolyStrata. One-click install for Windows and Linux, OpenAI/Anthropic API on localhost.

Llmfit helps founders find compatible AI models that can run on their specific hardware with a single command, simplifying the process of selecting appropriate models for their projects.

Sno Station — your Claude Code and Codex working as one squad on your own machine. Shared encrypted memory, agent-to-agent messaging (Reach), squad skills for handoff and cross-vendor review, and a nightly loop that rewrites the agents' own skills with your approval. Open source, no daemon, no cloud required. Assembled in public.

Llama.cpp enables fast and efficient inference of large language models in C/C++ applications, allowing developers to integrate AI capabilities into their projects

Stirling-PDF is a TypeScript-based application that allows users to edit PDFs on any device, anywhere, making it a versatile tool for managing and modifying PDF documents.

Llmfit helps founders find compatible AI models that can run on their specific hardware with a single command, simplifying the process of selecting appropriate models for their projects.

Koreader is an ebook reader application that supports various formats such as PDF, DjVu, EPUB, and FB2, allowing users to read ebooks on devices like Cervantes, Kindle, Kobo, PocketBook, and Android devices

MNN is a high-performance, lightweight inference engine for on-device AI and Edge AI applications, allowing for fast and efficient model deployment

Runanywhere-sdks allows you to run AI models directly on your device, eliminating the need for cloud services and enhancing data privacy and security

Enables running Armbian on various Amlogic, Allwinner, and Rockchip devices, including those with specific chipsets like a311d, s922x, and rk3399

GenieX allows startup founders to run advanced AI models like LLMs and VLMs directly on Qualcomm devices using local processing units such as NPU, GPU, and CPU

Tutanota encrypts emails, contacts, and calendar entries to provide secure communication for founders and their teams.

rtl_433 decodes radio transmissions from devices on the ISM bands, allowing you to capture and interpret data from IoT devices and sensors operating at 433mhz

A plain‑text, Markdown‑based note‑taking and todo‑list app that syncs your files with Nextcloud/ownCloud.

Maka-agent provides a local-first AI desktop assistant for startup founders to interact with their computer using natural language commands.

Openmed is a local-first solution that performs clinical named entity recognition (NER) and HIPAA protected health information (PHI) de-identification on-device, without sending patient data to the cloud.

Codex-keysmith enables version-independent deployment of Codex instructions with features like dry-run, backups, hook isolation, and recovery for reliable AI model management.

OpenBiliClaw is a private AI agent that provides personalized content recommendations from various platforms like Bilibili, Xiaohongshu, Douyin, and YouTube, based on your usage, feedback, and dialogue.

LiteRT deploys high-performance machine learning models on edge platforms, enabling efficient and optimized on-device deployment of ML and GenAI models

OpenUI5 helps you build responsive web applications that run on almost any browser and are accessible on various devices

Eidos lets you keep a relational spreadsheet in a single TypeScript file that both you and AI agents can read and write, turning flat CSV‑like data into linked tables you can query programmatically.

Club-3090 provides community-driven recipes for serving large language models on NVIDIA RTX 3090, 4090, and 5090 CUDA GPUs, supporting multiple engines and models

Auto‑Round quantizes large language models to low‑bit (e.g., 4‑bit) weights while keeping accuracy comparable to full‑precision inference.

Run any local LLM engine, auto-tuned to your GPU — polished web UI + OpenAI/Anthropic-compatible API. Point Claude Code at your own machine in one command. No Electron, no Python, offline-first.