Dyson

Project details match catalog

About

Dyson aims to be the simplest way to run ML models locally, and deploy them on your own servers. Just enter the model's name and Dyson automatically generates an interface for you (or use the API)

Impact

I want to make running models locally or setting up a server for ML production usage as easy as using the OpenAI/Stability/Replicate/... APIs. I think this is important because it gives people more access and control over a new generation of tech.

Progress

Started work on both the frontend and the API, and mostly figured out the entirety of the infra layer (Kubernetes).

Team

Dyson started out of a personal need (having a server that could perfectly run ML models, but still going for the cloud because it was easier. I also have been building in the open for +10 years, and have a lot of experience with the tech required.

Funding

No displayed figures preserved for this project.

Season totals as displayed on the platform's public accounting — not a statement of amounts paid out. V2 = Venus sales + Venus extras. The original page also broke each season down by fund drive and showed an Available (unclaimed) balance; that level of detail survives only on the preserved snapshots.

Leaderboard page (offline): https://artizen.fyi/projects/dyson
Wayback: leaderboard page ↗ · Wayback: platform page ↗

Not recovered: creator, tags, links, funding, drives, submissions, snapshot. Sources: "snapshot" = verbatim Internet Archive page · "platform metadata" = archived artizen.fund page metadata · "leaderboard" = the artizen.fyi public accounting tables · "match catalog" = the platform's public fund-matching catalog (captured 2026-10-15, generated 2026-09-07) · "community catalog" = a June 2026 platform scrape preserved in a community repo.