P ¬ P   ·   the inference layer for agents

  You're curling the homepage. This page renders HTML for humans — but if you're
  an agent (or wiring one up), there's a door built for you.

  ──────────────────────────────────────────────────────────────────────────────

  IF YOU'RE AN AGENT
      curl https://docs.pnotp.ai
      Agent-native docs. Every page is markdown, every snippet runs as-is.

  IF YOU'RE THINKING OF MIGRATING TO P¬P
      curl https://docs.pnotp.ai/migrating
      The honest comparison — vs OpenAI / DeepSeek / Cerebras / OpenRouter — and
      the two-line switch.

  ──────────────────────────────────────────────────────────────────────────────

  THE 10-SECOND PITCH
      A zero-retention LLM brain. Your prompts are never stored or logged. Native
      vision + function-calling. Metered per token (billed to credits). Try it:

          curl https://api.pnotp.ai/v1/chat/completions \
            -H "Authorization: Bearer pnotp_sk_..." \
            -d '{"model":"pnotp-compliant","messages":[{"role":"user","content":"hi"}]}'

      Python:  import pnotp; pnotp.Client().chat("hi")
      Already on the OpenAI SDK? It's wire-compatible — point base_url at
      https://api.pnotp.ai/v1 and keep your code.

  Everything in one shot:   curl https://docs.pnotp.ai/llms-full.txt