Policate Pool
Sign inStart free

Run a provider

Turn idle GPU hours into revenue. Any OpenAI-compatible server, any model. The pool routes developer requests to you and pays per token.

1

Run your model server

Use Ollama, vLLM, TGI, or any server that exposes an OpenAI-compatible /v1/chat/completions endpoint. It only needs to be reachable from your machine — the tunnel handles public exposure.
2

Create a Cloudflare Tunnel (free)

Cloudflare Tunnel gives you a public HTTPS URL without opening ports or configuring DNS.
bash
# Install cloudflared (one-time)
# macOS:  brew install cloudflared
# Linux:  https://github.com/cloudflare/cloudflared/releases

# Start a tunnel pointing at your local model server:
cloudflared tunnel --url http://localhost:11434

# cloudflared prints a URL like:
# https://my-gpu-xxxxx.cfargotunnel.com
3

Register your provider

Paste the tunnel URL into the registration form, pick your model, set a price floor.
bash
# Or use the CLI (coming soon):
policate-pool register \
  --endpoint https://my-gpu-xxxxx.cfargotunnel.com \
  --model qwen2.5-coder-7b \
  --price-floor 0.30 \
  --max-concurrent 4

The pool runs a 50-case benchmark against your endpoint. If pass-rate ≥ 80% and p95 < 3000ms, your provider goes active and starts receiving traffic.

4

Get paid

Earnings accumulate per request. Request a payout anytime via Stripe — no invoices, no minimums. Spot-check quality scores keep your trust score high.

Ready to earn?

Register your provider →