iwant
Product Hunt Oskar Kocol
iwant sets up your own AI model server with one command. It searches GCP for free GPUs, then handles the messy setup for you.
Based on reporting by Product Hunt, Oskar Kocol — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
iwant is a command-line tool built to make self-hosting easier for people who want to run their own model endpoint instead of renting one from someone else. From a single command, it can spin up support for models like DeepSeek, Kimi, Minimax, MiMo, and Step.
The part that does the heavy lifting is pretty bluntly practical. iwant hunts across Google Cloud regions for available GPUs, then takes care of provisioning, downloading the model, and wiring up vLLM. That means the user doesn’t have to stitch together each step by hand.
Once it’s done, the tool hands back an OpenAI-compatible endpoint, an API key, and SSH access. That’s a neat package for anyone who wants to test or run models without living in cloud-console purgatory.
The launch page lists iwant as an open source developer tool in the AI infrastructure category. So the pitch isn’t mystery or magic; it’s removing friction from the very unglamorous work of getting a model online and usable.
My take — AI-written commentary, not fact-checked reporting
This is the right kind of open-source AI tooling: boring in the best way, because the boring bits are what stop people from actually using open models. The industry keeps pretending that model access is the hard part, when half the battle is just getting GPUs, endpoints, and auth to behave.
Read more about this at: Product Hunt