Unplugg
Forecasting API for timeseries data.
About Unplugg
Unplugg provides an API for forecasting timeseries data, such as energy consumption, temperature, or other data with seasonality effects. Access requires an API key/access token, and requests are made via POST to a /forecast endpoint at api.unplu.gg, with authentication via an x-access-token header.
The API accepts a JSON payload containing timestamped data points (Unix time format), an optional forecast_to date specifying how far to forecast, and an optional webhook callback URL for receiving results. If the submitted timeseries is not equally spaced, the API resamples it based on the interval between the first two data points. Malformed payloads return a 422 error with a descriptive message.
Upon processing, the API returns a job_id, and results are delivered to the specified webhook as a JSON object containing the forecasted values and the associated job_id. If no forecast_to date is provided, the forecasted array defaults to a length of one-eighth of the input timeseries length. The service is built by Whitesmith and includes documented example use cases for temperature prediction, electricity consumption, and ISP traffic forecasting.
Key features
- Forecasts timeseries data such as energy consumption, temperature, or ISP traffic
- REST API with a single POST /forecast endpoint
- Automatic resampling of unevenly spaced timeseries data
- Asynchronous processing with webhook callback delivering results
- Configurable forecast end date via forecast_to parameter
- Interactive online demo to test forecasts before integrating the API
Frequently asked questions
What does the Unplugg API do?
It provides forecasting for timeseries data with seasonality, such as energy consumption or temperature. You send historic data and receive a forecasted continuation of the series.
How do I authenticate requests?
You request an access token and include it in the x-access-token header along with Content-Type: application/json on requests to the API.
How is the forecast result delivered?
You provide a callback URL in the payload, and once the forecasting job finishes, the result is POSTed to that webhook as JSON containing the forecast array and job_id.
What format does the timeseries data need to be in?
Data is a list of objects with a Unix timestamp (seconds) and a numeric value; unevenly spaced series are automatically resampled.
What happens if I send a malformed payload?
The API returns a 422 UNPROCESSABLE ENTITY status along with a message to help locate the error.
Show your product to thousands of developers
· 100k monthly pageviews
· 7k newsletter subscribers
You might also like
Roboflow Universe
AIPre-trained computer vision models.
SkyBiometry
AIFace Detection, Face Recognition and Face Grouping.
Spam Hunter
AIFree service to classify text as spam using ML.
Summarize Text with AI
AIThis API generates customizable summaries of text and web pages using AI.
TranscriptFetch
AIYouTube transcript API for developers. Fetch captions from any video in production—no proxies, no IP blocks. Built for LLM, RAG, and MCP pipelines.
WolframAlpha
AIProvides specific answers to questions using data and algorithms.
