# Purpose-built for energy market forecasting

## From Live Data to Production Forecasts, Automatically

For in-house forecasting teams who want their own models running on live data — pd4castr watches your data sources, runs your models when inputs are ready, and delivers production forecasts your team can use right away.

### 100% Reproducible runs
### Zero Custom pipelines

## How pd4castr Works

From local model authoring to production forecasts — one continuous flow

```yaml
pd4castr.config.yaml

inputs:
  - aemo_predispatch
  - demand_5min
trigger:
  on: input_ready
```

### Author your model

Define inputs, outputs, sensitivities, and trigger behavior in one configuration, then package the model in Docker. Test locally with the same I/O contract used in production so promotion to cloud execution is predictable.

**Inputs**  
- AEMO predispatch  
- Live demand  
- Wind forecast  
- NEMDE residuals

3 of 4 ready

### Automated data ingestion

pd4castr continuously monitors upstream data sources, tracks input readiness, and ensures your models always have complete, current data.

```plaintext
price-forecast / r3
```

### Validate inputs

Resolve dependencies

Compute forecast

### Event-driven execution

When your data is ready, models execute automatically with secure, credential-free I/O in the cloud. No manual triggers, no scheduled guesswork.

**$/MWh forecast**

### Results delivery

Results are available immediately through APIs and dashboards. Every run is traceable and comparable, ready for your trading, risk, or planning teams.

## Why pd4castr

Move faster from research to production — without building the infrastructure in between

### Automated Data to Forecast

pd4castr monitors your data sources and tracks input readiness, so forecasts trigger when inputs are truly complete — not when a cron fires and hopes for the best.

### Reproducible by Default

Every run captures its exact inputs and outputs — traceable, auditable, and easy to compare across versions and time.

### Same Model, Any Environment

Run locally for iteration or in the cloud for production. Same interface, no environment-specific rewrites.

### One Platform, Many Models

Support multiple models, scenarios, and delivery paths without building a custom pipeline for each one.

## Built for Every Team

pd4castr delivers value across your entire forecasting organization

### ML / Data Scientists
- Run experiments and sensitivities without building the infrastructure around them
- Compare runs reliably and iterate locally
- Deploy your model to production without rewriting it
- Focus on modeling, not infrastructure

### Engineers / Platform Teams
- Replace bespoke schedulers and fragile glue code
- Reduce operational risk with a consistent, repeatable process
- Get traceable run history and clear auditability
- Deploy and scale without environment-specific rewrites

### Managers / Leadership
- Shorten time-to-value for forecasting initiatives
- Reduce operational overhead and engineering burden
- Get more dependable forecasts with clear auditability
- Enable faster scenario planning for commercial decisions

## Ready to Transform Your Forecasting Workflow?

Sign up to your first forecast run in minutes — no pipelines to build, no infrastructure to manage.

## Platform at a Glance

Everything you need for production forecasting

### Event-Driven Execution

Trigger runs as soon as data arrives, not on brittle schedules

### Flexible Trigger Rules

Configure each input to wait for fresh data or use the latest available — so triggers match how your data actually arrives

### Container-Based Model Runs

Package models cleanly and run them consistently across environments

### Secure Model I/O

Models don't need cloud credentials — inputs and outputs are handled securely by the platform

### Forecast Traceability

See what ran, when, with what inputs, and whether it completed successfully

### Scenario & Sensitivity Runs

Run what-if scenarios and sensitivities alongside your baseline, with comparable results out of the box

### Multi-Format Outputs

Ingest JSON, CSV, and Parquet outputs so teams aren't locked into one export style

### SDK Integration

Integrate forecasts into trading, risk, revenue, and planning systems

### Interactive UI for Results

Visualize forecasts, compare runs, and review key metrics without custom dashboards