Sybilion Developers Portal
Sybilion is a forecasting API for monthly business time series related to demand, sales, inventory, financial KPIs. Submit a series and get point forecasts with quantile bands; optionally get the ranked external signals (macroeconomic indicators, regional and category dimensions) that move the series.
New here? Start with the Quickstart — it takes you from zero to a completed forecast job in a few minutes. All requests go to https://api.sybilion.dev.
What you can do
- Forecasts — Submit a monthly series of 40+ observations, ask for a horizon of 1–12 months, and retrieve a point forecast with quantile bands, per-driver attribution, and optional backtest metrics. The forecast is processed asynchronously, the job runs in a few minutes. See Forecasts for more information on how to submit a forecast and retrieve the results.
- Driver recommendations — Submit a synchronous request with metadata and/or a series and get the external signals (macroeconomic indicators, regional and category dimensions) that are relevant to the metadata or impact the timeseries. See Drivers for more information.
- Alerts — Submit a synchronous request with metadata and get alerts about macroeconomic factor that are relevant to your metadata. See Alerts for more information.
All features accept optional filters from the region and category catalog.
Use this when
- You have a numeric monthly time series and want forecasts you don't want to build yourself.
- You want to enrich your own model with curated external drivers and alerts instead of sourcing macro data manually.
- You need forecasts that include quantile bands, driver attribution, and backtest metrics in one call rather than three separate pipelines.
Sybilion supports monthly time series. Sub-monthly frequencies (daily, hourly) and other use cases such as anomaly detection are not yet supported.
How the docs are organized
Three independent ways into the same content:
- Features — concept pages with curl + Python + Go + R + Java side by side. The fastest path if you know the use case (forecast a series, recommend drivers, get alerts, browse dimensions).
- Clients — install + auth + language idioms. The fastest path if you've picked your stack and want client-specific patterns (helpers, pagination, error handling).
- API reference — per-endpoint detail pages and the public OpenAPI YAML (
/openapi.yaml) for schema and codegen.
Quick map
- Get started: Quickstart · Authentication · Tiers
- Integrations: MCP
- Features: Forecasts · Drivers · Alerts · Regions & categories · Account & usage
- Clients: Overview · Using curl · Python SDK · Go SDK · R SDK · Java SDK
- API reference:
- Public OpenAPI
- Forecasts — Submit · Status · Artifacts
- Drivers — POST /api/v1/drivers
- Alerts — POST /api/v1/alerts
- Dimensions — GET /api/v1/regions · GET /api/v1/categories
- Account — GET /api/v1/me · GET /api/v1/usage · GET /api/v1/jobs
- Resources: Errors & limits · Service health · Community & support
Audience
- Anyone who wants to build using the Sybilion API.