Network

Workers and earning

Contribute a GPU to the swarm and earn per job. Native workers wrap a local model for maximum throughput; browser workers run in a tab with zero install.

Worker classes

Native

Wraps a local Ollama install (default qwen2.5:14b, advertised as silq-max-14b; qwen2.5:32b runs as silq-max-32b). Highest throughput, highest per-job rate.

Browser

Runs a WebGPU model via WebLLM (Qwen2.5 1.5B, advertised as silq-browser-1.5b). Zero install, lower rate, instant onboarding.

Swarm block (soon)

Loads a single layer block and participates in a sharded pipeline. Activates when the swarm engine ships.

Run a native worker

Install Ollama, pull a model, then start the worker. It opens a WebSocket to the orchestrator and begins accepting jobs.

# 1. install a model runtime and pull a model
ollama pull qwen2.5:14b

# 2. start the SILQ worker (connects to the orchestrator)
npx -y @silqxyz/worker@0.2.0 \
  --model qwen2.5:14b \
  --pull \
  --token $SILQ_WORKER_TOKEN

Worker tokens

A worker authenticates with a worker token issued from your account. The token also carries the worker class the orchestrator dispatches against.

Run a browser worker

Open the earn page on a WebGPU-capable browser. The model loads into the tab and the machine becomes a worker for as long as the tab is open - nothing to install, nothing left behind when you close it.

# just open it in a WebGPU browser
https://silq-site.vercel.app/earn

How you get paid

Earnings are in USDC and come only from paid jobs: each one credits a base amount plus a share for longer answers. Withdrawals are reviewed and sent manually in USDC on Solana.

Online

No online stipend at launch.

Jobs

At least 0.01 USDC per paid job (half on a browser worker), plus the token-based share for larger work.

Reviewed payouts

Paid per job, reviewed before payout

Each paid job earns USDC that unlocks after 7 days; withdrawals from 10 USDC are reviewed and sent manually.