Data Source
sales_2024.csvConnected
42K rows·12 columns·Daily
Anomaly Detection

Statistical outlier detection with configurable sensitivity. Auto-flags unusual patterns in your data stream.

Forecast Studio

Visual, node-based time-series forecasting. Connect data, build pipelines, and deploy predictions — all without code.

Forecast Engine
Prophet + XGBoost ensemble · 90-day horizon · Weekly seasonality · Holiday effects enabled
3 models · Cross-validated
Running
Start building
Dashboard
90-day forecast
Alert
Anomaly detectedWarning

Revenue dropped 23% below forecast on Mar 12. Triggered Slack notification.

Alex
Sarah
The platform

Visual forecasting on a canvas.

Drag nodes, wire pipelines, run predictions. Powered by Google TimesFM 2.5 — no data science required.

  • Connect any data

    Upload CSV, hit a REST endpoint, or connect Postgres. Forecast handles ingestion, schema inference, and basic cleaning automatically.

  • TimesFM 2.5 ensemble

    Google's foundation model + Prophet + XGBoost ensemble out of the box. Works on weekly seasonality, holiday effects, and bolts onto your data.

  • Live dashboards

    Forecasts render with confidence intervals as soon as the run completes. Embed via iframe or export to PDF.

  • Branching pipelines

    Compare model variants and feature engineering side by side. Promote the winner with one click.

  • Realtime + scheduled

    Run forecasts on demand or on a cron. Anomaly alerts trigger Slack / email / webhook the moment thresholds break.

  • Anomaly detection

    Statistical outlier detection with configurable sensitivity. Auto-flag drift before it becomes a problem.

How it works

Three steps from data to forecast.

  1. Connect data

    Drop a CSV, point at an API, or link a database. Forecast infers the schema and the time index automatically.

  2. Wire the pipeline

    Drag transform / model / alert nodes onto the canvas. Sensible defaults run out of the gate; tune when you want.

  3. Run + monitor

    Forecasts render with confidence intervals. Schedule reruns, set alerts, and embed the live dashboard wherever it's needed.

Built for

Who uses Forecast.

  • Operations

    Inventory + capacity in one place.

    • Demand forecasts for procurement
    • Capacity planning across locations
    • Drift alerts the moment patterns change
  • Finance

    Revenue + cash forecasts that update themselves.

    • Rolling 90-day revenue projections
    • Cash position with seasonal effects
    • Variance analysis vs. plan
  • Growth

    Acquisition + retention forecasts.

    • Cohort projections by channel
    • LTV with seasonality
    • Churn drift alerts
FAQ

Frequently asked questions

  • Do I need to know data science?
    No. Forecast ships with Prophet + XGBoost + TimesFM 2.5 ensembles tuned for the most common business series. Defaults work; tune when you want.
  • What data sources are supported?
    CSV upload, REST endpoints, Postgres, MySQL, BigQuery, and Snowflake. Custom sources via the Python SDK.
  • How accurate are the forecasts?
    On standard business series (revenue, traffic, inventory) the ensemble typically lands within 5-10% MAPE on a 90-day horizon. The dashboard shows calibration vs. actuals so you always know where you stand.
  • Can I embed forecasts elsewhere?
    Yes — embed via iframe, query the JSON API, or push to a webhook on every refresh.
  • What about anomaly detection?
    Drop an Alert node into your pipeline. Configure thresholds, anomaly sensitivity, or forecast drift limits. Notifications go to Slack, email, or webhooks automatically.

Forecast what's next.

Start building free. No credit card required.

Start freeRead the docs