Hey, it's Cindy 馃尡 You commented vibe, so here are all five from the reel with the real links, plus what each one needs to run properly.
btw the waitlist for AI System Club is now open! If you're keen on using AI to automate your content, video editing, job search, business or build AI agents, feel free to sign up (only if you want to hehe). Ps. the waitlist gets first dibs:)
the five from the reel
Over 688,000 stars between them
All five run on your own laptop. Nothing uploaded, no per-use cost, no API key. The catch is hardware, and each one below says what it actually needs.
1 路 OpenCode 路 211k stars An open source coding agent that runs in your terminal. The closest free thing to Claude Code. opencode.ai Repo moved: it is anomalyco/opencode now, not sst/opencode.
2 路 Ollama 路 182k stars Run Qwen, DeepSeek, Gemma and gpt-oss locally and offline. ollama.com A 7B model wants about 8GB of RAM. Smaller models run on almost anything.
3 路 ComfyUI 路 135k stars Node-based image and video generation, fully local. Your own studio with no credits to buy. Comfy-Org/ComfyUI Repo moved: Comfy-Org/ComfyUI, not comfyanonymous. A GPU helps enormously here.
4 路 Whisper 路 109k stars OpenAI's speech recognition, free and offline. Transcribe anything without uploading it. try it in the browser first
5 路 Upscayl 路 50k stars AI image upscaling as a normal desktop app, on Mac, Windows and Linux. upscayl.org
鈿狅笍 Two of these repos changed owner recently. OpenCode is now anomalyco/opencode and ComfyUI is now Comfy-Org/ComfyUI. The old links redirect today, but bookmark the new ones.
start here
The easiest two to try tonight
If you have never run anything locally, these two take about five minutes each and need no setup knowledge.
1 路 Upscayl, because it is just an app Download it, drag a photo in, done. No terminal, no model downloads, no configuration. upscayl.org
2 路 Whisper in the browser before you install anything Run it as a hosted demo first to see whether the quality is good enough for you, then install locally if it is. the hosted version
3 路 Then Ollama, which is one install and one command Install it, then pull a small model first. Start small and go bigger only if your machine copes. ollama.com
馃尡 Local models are not as good as the frontier ones, and that is fine. They are for the work you do not want leaving your machine, and for the work you do so often that per-use cost adds up.
the honest bit
What actually stops people
1 路 RAM, not skill Most give-ups are a model too big for the machine. If it crawls, pull a smaller one rather than concluding local AI is bad. model sizes
2 路 ComfyUI is genuinely steep It is the most capable and the hardest of the five. Do not start here. Come back once something else has warmed you up. Comfy-Org/ComfyUI
the waitlist is now open
AI System Club is launching soon
Build AI systems with me through recorded lessons you can follow at your own pace. Join the waitlist for first access and the founding price.
btw this list gets first dibs the moment doors open :)