Your machines. One swarm. The right model for every task.

llm-swarm is a local-first mixture-of-models host: one intake model streams the answer while specialist models fan out in parallel — routed by live queue depth and simple if/then rules. Join other machines as an opt-in collective with invite-based trust.

Download the client

Invites are single-use tokens minted by a host. Sign in with your invite email + Access OTP. Hosting your own swarm? See the host app docs.