NEW: Uncensored Llama-3 8B Stateful Engine

The Stateful & Uncensored LLM API for Games and Interactive Bots

Zero alignment filters. Managed long-term character memory. Ultra-low latency inference via a single HTTP POST request.

View Integration Docs
No history payloadPersistent character memorySub-100ms TTFT
import { CornAPI } from "@cornapi/sdk";

const corn = new CornAPI({ apiKey: process.env.CORN_API_KEY });

const reply = await corn.chat({
  character_id: "npc_mira_tavern",
  session_id: "player_42",
  message: "What happened here last night?"
});

// Memory + persona persist automatically.
// You never send the chat history.

Why CornAPI

Built for game loops, not chat UIs

One POST. A living character. No prompt archaeology, no refusal loops, no session stitching.

Managed Memory

We handle KV-cache and long-term RAG memory so your characters never forget context.

Zero Censorship

Unfiltered models fine-tuned specifically for creative writing, erogames, and roleplay.

Edge Latency

Sub-100ms time-to-first-token powered by optimized vLLM clusters.

Pricing

Simple plans for indie to studio

Checkout is gated while we onboard developers in batches. Selecting a paid plan still records purchase intent.

Indie Dev

Beta

$0/mo

  • 1,000 free requests/mo
  • Shared GPU queue
  • Community Discord
Most popular

Pro Creator

$10/mo

  • 50,000 requests/mo
  • Priority GPU queue
  • Persistent character memory

Studio

Dedicated

Custom

  • Dedicated RunPod instances
  • Custom SLAs
  • Private model routing