A self-hosted data logger for your Tesla 🚘 [main maintainer=@JakobLichterfeld]
Tesla Mod — Flipper Zero & ESP32. Nag killer, FSD region unlock, track mode, BMS dashboard, blind spot alert, high beam strobe, CAN capture, steering mode + more. Works via OBD-II or X179 connector.
Play your Nintendo Switch or Switch Lite on your PC: stream games at up to 1080p60 with sound over USB-C or the network, no dock or capture card. Record MP4, take native-resolution screenshots, or use it as a webcam. Atmosphère sysmodule + viewers for Windows, Linux and macOS.
Semantic search over videos using Gemini Embedding 2 or Qwen3-VL.
—
Tailscale for the Nintendo Switch: an Atmosphère sysmodule that puts homebrew (like Moonlight) on your tailnet
Make complex ideas click without losing accuracy or professional depth.
The flexible HTTP client library for Elixir, with support for middleware and multiple adapters.
Tesla custom integration for Home Assistant. This requires a refresh token be generated by third-party apps to login.
—
TeslaLogger is a self hosted data logger for your Tesla Model S/3/X/Y. Actually it supports RaspberryPi 3B, 3B+, 4B, Docker and Synology NAS.
Hand-written NVFP4 W4A16 CUDA kernels for Volta
Load test HTTP, database, and Redis on one clock and correlate their latency to pinpoint the real bottleneck — with a shareable HTML report
Turn vague product or feature ideas into clear, checkable requirements before planning or coding—mapping rough descriptions to professional terms, exposing ambiguities, and aligning goals, flows, and boundaries. 这是一个需求澄清 Skill:把说不清的想法整理成一份清晰、可核对的需求清单,在执行任务前对齐目标、流程与边界等。
The set-and-forget LLM engine for Pascal and Volta: PXQ codec + kernels, auto-tuned per card. Ready-to-run PXQ models in MODELS.md; benchmarks vs llama.cpp in the README.
Tesla multi-intent in-car agent (RAG + Skills)
Run Tesla OS On QEMU
Your Tesla in the Omarchy bar: where it is, how full it is, and how far that gets you. Without keeping the car awake.
Supervising whether AI really promotes cross-platform skills in research programs
AI skill for typed, auditable workflow graphs, conditional routing, evidence-gated evolution, and validation. Includes full source, examples, and tests.
llama.cpp performance patches developed and measured on Tesla P100 (GP100, sm_60). Most are not Pascal-specific.
NInfer on Tesla V100 (sm_70): Qwen3.8-27B deployment, build/convert scripts, systemd service, and cross-engine benchmark (vs llama.cpp). Fork of Neroued/ninfer with sm70 vision-encoding speedup.
llama.cpp optimized for NVIDIA Tesla V100 (SM70): large MoE Qwen3.8-Flash-Next with MTP (prefill 3464 tok/s at 200K context, 6x V100-16GB) and Qwen3.8-27B with DFlash2 (decode up to 225 tok/s); 2-6 GPUs, 262K-512K context, multimodal, concurrent serving.
鹿路录 RUCareer:本地优先的个人经历档案、AI 辅助简历制作与官方下载网站
代码解读器:把代码、脚本、程序、仓库与 AI skill 转述为易懂的自然语言,讲清实际流程、数据变化与关键分支。
破茧:YouTube 可解释反推荐浏览器扩展
A cross-agent, reality-first workflow for mathematical modeling
Agentic-coding oriented inference engine for N x V100-SXM2, based on SGLang: GLM-5.3-Flash, Qwen3.8-Flash-Next, DeepSeek-V4.1-Flash, MiniMax-H3.
ComfyUI MiniMax H3 inference optimizations for NVIDIA Tesla V100 (SM70)
tesla-media-hub 的 Cloudflare Worker 版:前端静态资源 + AppleCMS 点播 API(已移除 IPTV/ffmpeg),支持 Git 集成一键部署
An excerpt from our financial valuation model of Tesla
sunnypilot is an open source driver assistance system. sunnypilot offers the user a unique driving experience for over 350 supported car makes and models with modified behaviors of driving assist engagements. sunnypilot complies with the safety policy from comma.ai's openpilot as accurately as possible.