Skip to main content
A component is a self-contained piece of analysis with typed inputs and outputs. Components are the things experiments run; they’re also what the dataset’s Components tab assigns and configures. You don’t write components — they’re built by Summand and exposed through a catalog the in-product editor reads to render forms, validate inputs, and dispatch runs.

​
The catalog

These components are visible in the experiment editor:

Column stats

Per-column distributions, missingness, cardinality, and basic numeric statistics.

Predictors (EBM)

Fits an Explainable Boosting Machine and writes feature importance, shape functions, and pairwise interactions.

Correlation-association matrix

Correlations and other measures of association for every column pair. Beta

Forecast

Time-series projection with 95% prediction intervals, using a Holt-Winters model under the hood. Beta

Outlier report

Identifies unusual values and anomalous rows. Beta

Feature metadata

LLM-generated, business-friendly descriptions for every dataset column. Powers chat grounding.

UMAP embedding

A 2-D projection of feature space for cluster visualization and “similar rows” queries.

Missing-value patterns

Identifies patterns in the missing data. Beta

Distribution profile

Per-column information about the shape of the data distribution. Beta
The editor also lists EBM prediction intervals (ebm_calibration), which adds prediction intervals to a trained EBM, and Batch prediction (score_dataset), which scores every row of a dataset with a trained model and writes the results as a new scored dataset. Other components (ebm_model, forecast_model, and the XGBoost, LightGBM, CatBoost and linear model trainers) exist behind the scenes. They hold the heavy fitted state and are consumed by the user-visible ones. They’re hidden from the editor; ebm_model and forecast_model run automatically when a component that depends on them is selected.

​
Anatomy of a component

Every component declares the same metadata: When you select a component in the experiment editor, the catalog is what tells the form which inputs to render and how to validate them.

​
How components run

Components within a wave run in parallel; waves run sequentially because of dependencies.

​
Reading component outputs

Two places where outputs surface:
  • Dataset detail → Components tab. Per-component status plus the latest artifact, rendered with either the component’s declared blocks or a bespoke React viewer (for EBM, UMAP, and Feature metadata).
  • Summand chat. The analyze tool resolves the latest version of any component and returns its data inline. Filtering by column or feature name is supported through each component’s agent_config.

​
What’s not a component

A few analyses run outside the component catalog:
  • Surprise finding has its own Step Functions pipeline that runs alongside the semantic-layer dispatcher. It’s not selectable in the experiment editor today — surprises surface through their own page in the product.
  • Sync / curation (the read-from-source-and-write-to-Parquet step) is part of the connector pipeline, not the component catalog.
  • Chat itself isn’t a component — it’s a separate Lambda (claude-stream) that consumes component outputs via the analyze tool.

​
Catalog stability

The catalog is small and slow-moving. Customers can’t add their own components. If you need an analysis the catalog doesn’t cover, email support@intelligible.ai and tell us what you’d build.

​
Per-component details