Celeste VD × Groq

The AI Dungeon Master,
at the speed of thought.

Celeste VD runs its built-in AI DM on Groq — near-Claude narration, a free tier, and answers in under a second.

Speed

Real-time, because the table is real-time.

A Dungeon Master that pauses to think breaks the scene. Groq's LPU inference returns full narration in a fraction of a second — the story keeps its rhythm.

< 1s

to a full DM reply

Quality & cost

Nearly Claude. For free.

gpt-oss-120b

An open model served on Groq — narration close to Claude for the everyday table.

Free tier

Runs at no cost within Groq's free limits — plenty for a home campaign.

Claude when you want the best

Celeste takes any key — Groq, OpenAI, Gemma, or Claude for the richest DM.

For developers

Drop-in. OpenAI-compatible.

Same API shape as OpenAI — swap the base URL, the key and the model. No rewrite. That's how Celeste plugs Groq into its DM engine.

import OpenAI from "openai";

const groq = new OpenAI({
  apiKey: process.env.GROQ_API_KEY,
  baseURL: "https://api.groq.com/openai/v1",
});

const res = await groq.chat.completions.create({
  model: "openai/gpt-oss-120b",
  messages: [{ role: "user", content: "Narrate the ambush." }],
});

Groq builds the fastest inference for AI. Celeste VD is proud to run its AI DM on it.

groq.com →