Sign in / Workspace
Back to the good stuff
TWO WAYS IN. SAME EXPRESSIVE FACES.

A signal now.
Or a signal that keeps coming.

Have the whole signal already? Send it over.
Still collecting it? Send one value at a time. We’ll keep the context.

Send the signal. Get the story.

For a recorded run, a chunk of readings, or a signal you’ve already collected.

  1. POST your samples in time order.

    Send an array of values in one request. The detector moves through the signal, looking for moments that break the pattern.

  2. Get a result for that signal.

    The response includes the overall verdict, anomaly locations, scores, and emoji severities. No drama? 😌, severity 0, and an empty anomaly list.

  3. Start fresh with the next request.

    Each request stands on its own. The local demo does not store your samples or remember the previous request.

Try the working playground
ONE REQUEST INIllustrative request / response
POST /v1/analyze
Authorization: Bearer YOUR_API_TOKEN
Content-Type: application/json

{
  "signal": [42, 41, 43, 42, 41, 43, 42, 85]
}
ONE RESULT OUT response excerpt
{
  "verdict": {
    "emoji": "😱",
    "severity": 8
  },
  "uhohs": [{
    "index": 7,
    "value": 85,
    "emoji": "😱",
    "severity": 8
  }]
}

Your signal keeps going. So do we.

For a sensor, a service metric, or anything with another reading on the way.

Instead of sending the entire history every time, start a stream and send each new value with its stream ID. We’ll keep a window’s worth of recent samples for that stream and use the updated window to analyze the newest value.

New valueUpdate windowAnalyze newest sample😌 → 😱
A WINDOW OF 5 SAMPLES

A little memory goes a long way.

INTERACTIVE ILLUSTRATION

Send the next value to see what stays, what leaves, and what gets a verdict.

OUT—Nothing yet
STORED WINDOW4 / 5
42sample 1
41sample 2
43sample 3
42sample 4
···waiting
NEXT IN41One new value
⏳
Warming up · 4 of 5 samples

One more sample before the first verdict.

Illustration only: the 85 spike is scripted to show an uh oh. This does not run a model or store data on the server.

01 / FILL

First, a little context.

Your model’s receptive field sets the context size. While it fills, the stream reports “warming up.” No verdict yet means “not enough data,” not severity 0.

02 / SLIDE

One in. One out.

Once full, each arrival replaces the oldest value. A model keeps its receptive field’s worth of prior context to predict the next sample.

03 / REACT

A fresh face each time.

Each new value gets its own result based on the current window. Yesterday’s uh oh doesn’t make every future reading an uh oh.

THE STREAM FLOW

Same signal. Same stream ID.

Keep different sensors or metrics in separate streams. Send their samples in time order so the stored window represents the right history.

Start sequence 0 and use next_sequence from each response. Retry the last sequence with the same values to replay its result. Use a new stream ID to start fresh, or DELETE the old stream. Send samples in order; timestamps are not inferred.

LIVE ENDPOINT · BEARER TOKEN REQUIRED
POST /v1/streams/sensor-01/samples
Authorization: Bearer YOUR_API_TOKEN
Content-Type: application/json

{"sequence": 0, "value": 41}

Use next_sequence from the response.
Fill the model context, then get fresh verdicts.
WHICH ONE FITS?

A snapshot or a running conversation.

One-time signalOngoing stream
You sendA complete array of samplesA stream ID and each new value
We keepNo history between requestsA rolling window per stream
You getA summary and anomalies across the signalA verdict for each new value after warm-up
Best forRecorded runs and collected readingsLive sensors and continuous metrics
Preview statusToken-based model inferencePersistent model history
🧠

What’s running today? The homepage uses a statistical demo. Authenticated API calls run your saved TCN: complete signals at /v1/analyze, or ongoing samples at /v1/streams/{stream_id}/samples. Create a model-specific token in your workspace’s API tab. Stream history persists in the database. The five-sample illustration above is explanatory; real warm-up depends on your model.

😬

Start with a little signal.

Meet the faces before you meet the API.

Back to the playground