We built a tool that installs frameworks like ComfyUI, Ollama, OpenWebUI etc on any cloud GPU in one command and saves your whole setup between sessions [R]
swm is an open-source tool that simplifies cloud GPU usage by installing frameworks like ComfyUI and Ollama in one command, and automatically saves your entire workspace between sessions, enabling seamless migration across providers.
We kept running into the same problem every time we rented a GPU to run Ollama + OpenWebUI or ComfyUI, we'd spend the first 45 minutes reinstalling everything. Custom nodes, models, configs, all of it. Docker images went stale fast, different providers had different base images, and nothing was truly portable. We got sick of it and built swm. Here's what it does for ComfyUI users specifically: swm gpus -g a100 --max-price 2.00 --sort price shows you the cheapest available GPU across RunPod, Vast ai, Lambda, and 7 other providers in one view swm pod create — spins up an instance on whatever provider you pick swm setup install comfyui — installs ComfyUI on the pod From there the main thing is the workspace sync. Your entire setup custom nodes, models, outputs, configs lives in S3-compatible object storage (I use B2). When you're done you run swm pod down and it pushes everything, kills the instance, and next time you spin up on any provider you just pull and everything is exactly where you left it. No more reinstalling 15 custom nodes and redownloading checkpoints every session. We also built a lifecycle guard because we kept falling asleep mid-session and waking up to dumb bills. It watches GPU utilization and if nothing's happening for 30 minutes (configurable), it saves your workspace and terminates automatically. Has saved us more money than we want to admit lol. A few other things: * Background auto-sync daemon pushes changes every 60 seconds so you don't have to remember to save * Tar mode for huge workspaces with tons of small files packs everything into one S3 object instead of 600k individual uploads * Also supports vLLM, Ollama, Open WebUI, SwarmUI, and Axolotl if you do more than SD * Works with Cursor, Claude Code, Codex, Windsurf if you want your AI agent to manage GPU instances for you Free, open source, Apache 2.0. pipx install swm-gpu Site:[ https://swmgpu.com](https://swmgpu.com/) GitHub:[ https://github.com/swm-gpu/swm](https://github.com/swm-gpu/swm) Would love feedback from anyone who rents GPUs. What's the most annoying part of your current workflow? We are also looking for contributors to the open source repo and suggestions on new frameworks/extensions to be included. Please share your thoughts
Floci is a newly open-sourced, lightweight AWS emulator that runs 45 services on a laptop using only 13 MiB of memory, offering a faster and resource-efficient alternative to Docker-based tools like LocalStack.
A developer built a local web GUI for the TradingAgents multi-agent LLM stock analysis framework, supporting various LLM providers and adding features like live pipeline visualization, a report reader, and multi-session chat.
AgentSwarms launched an interactive, gamified blog that helps users match open-source LLMs to the right GPU by calculating VRAM requirements based on model size and quantization, turning infrastructure planning into an engaging experience.
Open WebUI Desktop launches as a native app letting users run local LLMs or connect to remote servers without Docker or terminal setup, featuring offline operation, system-wide voice input, and floating chat overlay.
OvertChat is a simpler, polished self-hosted alternative to Open WebUI for local AI models, featuring single Docker compose setup, built-in web search, and Kokoro TTS, all MIT licensed.