milliseconds.ai demos
Add your API key

Examples are free. Your console test key covers custom inputs and live runs using your workspace’s quota.

Kept in memory for this page. Clear and reconnect after changing workspaces. Check your workspace.

Pasted keys are saved in this browser. Keys are sent only to this site’s proxy for personal requests.

Get a free key
Sign up

All demosVoice Turn

Voice Turn

Know when to answer. Know when to stop talking.

Examples are free. Bring a key for your own inputs. Get a free key

Performance & cost

This session

Ready to run
Delivery time · median
—
Live model compute
—
Live input tokens
0
Est. cost · USD
$0.00

Bypass examples to measure live inference. Uses your quota.

Measurement details · 0 successful requests
Successful requests
0

Run an example for free to see its measurements.

Model compute is the average per measured request. Delivery time is the median browser time for successful requests, including queue, network, and retries. Cached delivery speed is not inference speed. Model compute can exceed browser time when one request processes many inputs.

Tokens come from response headers. Inference costs use $0.04 per million input tokens, before plan credits. The estimate includes live and recorded tokens at the same rate; cached examples are not charged again. Missing usage is unavailable or a partial total (≥). Demo resets keep these totals; reload to start a new session.
Try it

Start a session and follow the turn decisions.

Build with the SDKTypeScript · Python · dm1

One request from this demo. Copy it into your app.

Uses your MS_API_KEY and quota. Keep the key on your server.

Install · set your key
npm install @cloudraker/milliseconds
export MS_API_KEY="your-api-key"
example.ts
import { DecisionMachine } from "@cloudraker/milliseconds";

// Run on your server; reads MS_API_KEY from the environment.
const dm = new DecisionMachine();

const { result, usage } = await dm.classify(
  [
    "set",
    "set a",
    "set a timer",
    "set a timer for",
    "set a timer for ten",
    "set a timer for ten minutes",
    "a",
    "a timer",
    "a timer for",
    "a timer for ten",
    "a timer for ten minutes",
    "timer",
    "timer for",
    "timer for ten",
    "timer for ten minutes",
    "for",
    "for ten",
    "for ten minutes",
    "ten",
    "ten minutes",
    "minutes",
    "what's",
    "what's the",
    "what's the weather",
    "what's the weather like",
    "what's the weather like in",
    "what's the weather like in Boston",
    "what's the weather like in Boston tomorrow",
    "the",
    "the weather",
    "the weather like",
    "the weather like in"
  ],
  {
    "set_timer": "Start a timer, alarm or countdown for a duration",
    "weather": "Ask about weather, temperature or forecast",
    "play_music": "Play a song, artist, genre or playlist",
    "lights": "Turn lights on, off, dim them, or change their color",
    "send_message": "Send a text or message to a person",
    "question": "A general knowledge or factual question",
    "chit_chat": "Greeting, small talk or thanks with no task",
    "incomplete": "Too little said so far to tell what the speaker wants"
  }
).withUsage();

console.log(result);
console.log({ inputTokens: usage.inputTokens, modelMs: usage.inferenceMs });

TypeScript documentationNode 22+: node --experimental-strip-types example.ts

Under the hood How it works & limitations

API calls/classify/yes-no

How it works

The model identifies intent as transcript fragments arrive. Local rules check that required details are present before responding. While a reply is playing, a separate check detects an interruption.

Demo scope

The microphone and speech are simulated with a scripted transcript. Nothing is recorded; only transcript text is sent. Intent confidence describes the model’s judgment, not certainty that a request is complete.

Inspired by the work of dabit3.