One key.
Every model. Your terminal.
Sign up, add credits, create a key, install baton — and you're working with Claude, GPT, Gemini, Grok and more from your own terminal. Nothing to build first.
curl -fsSL https://quantumencoding.io/install.sh | sh
export QAI_API_KEY=qai_k_your_key
baton tuiGet started
From sign-up to a working agent in four steps
Create a key
Dashboard → Keys. One qai_k_ key works in baton, the SDKs and your own apps.
Open Keys →baton
The program your key unlocks
baton is our command-line toolkit for working with AI agents. Install it on macOS, Linux or Windows, set your key, and every model on the platform is a command away.
An agent in your terminal
baton tui reads and edits your files, runs commands and works through a task with you — on any model your key reaches. Switch with /model.
Media from the command line
baton qai makes images, video, music and speech with the same key and the same credits.
A fleet when you need one
A work queue, scheduling and messaging for agents across your machines — one ledger, no server to run.
Signed builds for macOS (Apple silicon), Linux (x86_64) and Windows (x86_64). More about baton · All downloads
Or stay in the browser
Describe it. We build it.
Signed in, the same models run here: chat a project through, plan it, and watch agents build it in a cloud sandbox.
SDK
Building your own app? Same key.
Six SDKs over one API. Five lines to your first call, and switching models is changing a string.
npm install quantum-ai-sdkimport { QuantumClient, userMessage, responseText } from "quantum-ai-sdk";
const client = new QuantumClient("qai_k_your_key");
const response = await client.chat({
model: "claude-fable-5-1",
messages: [userMessage("Build me a REST API for a todo app")],
});
console.log(responseText(response));
// Cost: $0.003 — billed automaticallyProof
We build on it too
Every product below runs on the same API, the same billing, and the same SDK you'll use. This is our proof it works.
This site, signed in
Mission Control
The workspace behind quantumencoding.ai — sign in and it is already here, no separate install or account. An architect agent decomposes your project, assigns models, and deploys worker agents in parallel. Watch them read, write, test, and commit in real time, on the same API you call.
- · Multi-model orchestration — pick a different model per task
- · Full project workspace with file I/O, git, and shell access
- · Visual workflow builder, dependency graphs, live agent streams
- · Test API calls, compare models, prototype before you code
Coming soon
Cosmic Duck
A consumer AI app for everyone — chat, voice conversations, image generation, video, music, document processing, for desktop, Android and iOS. Built on the same API, credits and key, as proof the platform scales to consumer products.
- · Chat with any model — Claude, GPT, Grok, Gemini, GLM
- · Realtime voice, TTS, STT, voice cloning
- · Image, video, and music generation
- · Cross-platform — Tauri desktop, Android, iOS
Models
Every provider. You choose.
Run the same task across providers and compare results, speed, and cost. We route. You decide.
Plus ElevenLabs, HeyGen, Meshy, Mistral, and more. Full model list in the docs.
Pricing
Pay for what you use
No subscriptions. No seat fees. No minimum commitment. Per-token pricing, transparent on every request.
How it works
Buy credits
Top up your balance. Credits stay valid for 18 months from purchase. Every API call deducts the exact token cost.
Every request
Transparent cost
Every response includes the exact cost in the headers. No surprises, no hidden fees, no mystery bills.
Volume
Spend more, pay less margin
Past $50 a month, apply for the Developer tier and the margin halves to 12.5%. Enterprise goes lower still.
Get started
Sign up. Top up. Get a key. Install baton.
No subscription, no minimum. One key, one bill, every model — in your terminal, your code, or right here.