Quickstart

Push a signal, create a goal, and get your first forecast — in about five minutes.

This walkthrough uses the fastest path through the platform: goals. You push some data in, tell predictAI what you want forecast, and the platform handles discovery, model selection, training, and deployment for you.

Get your credentials

You need two values: an API token and a workspace ID.

In the predictAI app, open Workspace Settings → API Tokens and create a token. Copy the pa_live_... value — it's shown only once.
Your workspace ID is on the same settings page (you can also list your workspaces with GET /v1/workspace).

export API_BASE="https://api.predict.ai"
export TOKEN="pa_live_..."
export WORKSPACE_ID="..."

https://api.predict.ai works for every account. If your data lives in another region the API answers with a one-time redirect (the SDKs follow it automatically); the exact base URL is shown next to your API token in the app. See Base URL & versioning.

Push a signal

A signal is a named time series. Push a value to a key and the signal exists — no schema to define first:

curl -X POST "$API_BASE/v1/signal/push" \
  -H "Authorization: Bearer $TOKEN" \
  -H "X-Workspace-Id: $WORKSPACE_ID" \
  -H "Content-Type: application/json" \
  -d '{
    "key": "daily_sales",
    "value": 12841.50,
    "timestamp": "2026-07-15T00:00:00Z"
  }'

from predictai import PredictAI

client = PredictAI(token="pa_live_…", workspace_id="ws_…")

data = client.signals.post_signal_push(json={
    "key": "daily_sales",
    "value": 12841.5,
    "timestamp": "2026-07-15T00:00:00Z",
})

import { PredictAI } from "@predictai/sdk";

const client = new PredictAI({ token: "pa_live_…", workspaceId: "ws_…" });

const data = await client.signals.postSignalPush({
  json: {
    key: "daily_sales",
    value: 12841.5,
    timestamp: "2026-07-15T00:00:00Z",
  },
});

{ "status": "ok" }

In practice you'd push a backlog of history (one call per point, or use a bulk import for large datasets). A forecast needs history to learn from — a few hundred points is a good start.

Create a goal

A goal tells the platform what outcome you care about. Everything else — which signals drive it, which model fits best, when to retrain — is discovered automatically:

curl -X POST "$API_BASE/v1/goals" \
  -H "Authorization: Bearer $TOKEN" \
  -H "X-Workspace-Id: $WORKSPACE_ID" \
  -H "Content-Type: application/json" \
  -d '{
    "target_key": "daily_sales",
    "goal_type": "forecast_value"
  }'

from predictai import PredictAI

client = PredictAI(token="pa_live_…", workspace_id="ws_…")

data = client.goals.post_goals(json={
    "target_key": "daily_sales",
    "goal_type": "forecast_value",
})

import { PredictAI } from "@predictai/sdk";

const client = new PredictAI({ token: "pa_live_…", workspaceId: "ws_…" });

const data = await client.goals.postGoals({
  json: {
    target_key: "daily_sales",
    goal_type: "forecast_value",
  },
});

{
  "data": {
    "goal_id": "b7e9c2d4-…",
    "run_id": "a1f3…",
    "status": "queued",
    "credits_charged": 25
  }
}

The 201 response means a discovery run is queued. The platform now profiles your data, finds related signals, runs a model tournament, and trains a champion. Watch progress with GET /v1/goals/{goal_id} — or subscribe to the goal's realtime channel instead of polling.

Get the forecast

Once the run completes and a champion is serving, the goal has a forecast:

curl "$API_BASE/v1/goals/$GOAL_ID/forecast" \
  -H "Authorization: Bearer $TOKEN" \
  -H "X-Workspace-Id: $WORKSPACE_ID"

from predictai import PredictAI

client = PredictAI(token="pa_live_…", workspace_id="ws_…")

data = client.goals.get_goals_by_goal_id_forecast(GOAL_ID)

import { PredictAI } from "@predictai/sdk";

const client = new PredictAI({ token: "pa_live_…", workspaceId: "ws_…" });

const data = await client.goals.getGoalsByGoalIdForecast(GOAL_ID);

For production integrations, call the inference alias — it always routes to the goal's current champion, so the URL survives champion swaps and re-deployments:

curl -X POST "$API_BASE/v1/inference/goals/$GOAL_ID/infer-segment" \
  -H "Authorization: Bearer $TOKEN" \
  -H "X-Workspace-Id: $WORKSPACE_ID" \
  -H "Content-Type: application/json" \
  -d '{}'

from predictai import PredictAI

client = PredictAI(token="pa_live_…", workspace_id="ws_…")

data = client.inference.post_inference_goals_by_goal_id_infer_segment(GOAL_ID, json={})

import { PredictAI } from "@predictai/sdk";

const client = new PredictAI({ token: "pa_live_…", workspaceId: "ws_…" });

const data = await client.inference.postInferenceGoalsByGoalIdInferSegment(GOAL_ID, { json: {} });

{
  "predictions": [
    { "label": "daily_sales", "point": [13102.4, 13350.9, 13571.2] }
  ],
  "horizon": 3,
  "inference_id": "3f1a…",
  "latency_ms": 245.67
}

Go further

Want full control instead of autopilot? Build the loop yourself: create a model, set up a pipeline, deploy, and infer.
Connect a database or upload files with Ingestion.
Set up alerts and realtime events so the platform tells you when something changes.
Understand what things cost before you scale up.

Get your credentials

Push a signal

Create a goal

Get the forecast

Go further

On this page