Inference for everyone

GPU owners earn. Developers save.

Run your GPU in the Policate pool — get paid per token. Access the pool as a developer — 3–5× cheaper than cloud.

How it works

For providers

Bring your GPU, earn per token

  1. 1
    Run your model
    Ollama, vLLM, any OpenAI-compatible server
  2. 2
    Create a tunnel
    One cloudflared command — public HTTPS URL
  3. 3
    Register
    Set your price floor and concurrency limit
  4. 4
    Get paid
    Weekly Stripe payout — no invoices, no contracts
Read the guide →

For developers

Cheap inference, drop-in API

  1. 1
    Top up credits
    $20 gets you started — Stripe checkout
  2. 2
    Point at the pool
    One env var change — OpenAI-compatible
  3. 3
    Use any compatible tool
    Pi, Claude Code, any SDK, curl
  4. 4
    Pay per token
    No contracts, no minimums, no AWS account
Read the guide →

60-second quickstart

Drop-in replacement for any OpenAI-compatible client.

bash
# 1. Create account at providers.policate.marcllort.com/signup # 2. Add $20 credits (Stripe) # 3. Create an API key export OPENAI_BASE_URL="https://api.providers.policate.marcllort.com/v1" export OPENAI_API_KEY="pool_live_xxxxx" # Use with Pi / policate CLI: export POLICATE_GATEWAY_URL="https://api.providers.policate.marcllort.com/v1" export POLICATE_API_KEY="pool_live_xxxxx"

Start in minutes

No AWS account. No IAM. No contracts. Just an API key and credits.