API Reference

This page provides a comprehensive reference of all FeatureByte REST API endpoints, organized by domain. For end-to-end examples, see the Credit Default Tutorial and Store Sales Forecast Tutorial.

Authentication

The API supports two authentication methods:

• Token-based: Pass a Bearer token in the Authorization header
• Username/password: Authenticate via POST /auth/jwt/login

import featurebyte as fb

fb.use_profile("your_profile")
client = fb.Configurations().get_client()

Catalog Management

Method	Endpoint	Description
GET	/catalog	List catalogs (supports pagination and search)
GET	/catalog/{catalog_id}	Get catalog details
POST	/catalog	Create a new catalog

Feature Store Browsing

Browse the data warehouse hierarchy: feature stores, databases, schemas, tables, and columns.

Method	Endpoint	Description
GET	/feature_store	List feature stores
GET	/feature_store/{id}/database	List databases in a feature store
GET	/feature_store/{id}/schema	List schemas in a database
GET	/feature_store/{id}/table	List tables in a schema
GET	/feature_store/{id}/table_details	Get table details (columns, types)
GET	/feature_store/{id}/table_preview	Preview table data (with row limit)
GET	/feature_store/{id}/column	List columns for a table

Parameters for browsing endpoints: feature_store_id, database_name, schema_name, table_name, limit (for preview).

Table Registration & Analysis

Source Table Analysis

Detect the table type and validate columns before registration.

Method	Endpoint	Description
POST	/table/source_table_analysis	Analyze a source table (async)
GET	/table/source_table_analysis/{id}	Get analysis results
POST	/table/source_table_summary	Generate AI-powered table descriptions (async)
GET	/table/source_table_summary/{fqn}	Get generated summary

Table Registration

Register tables in a catalog. Each table type has specific required fields.

Method	Endpoint	Description
POST	/event_table	Register an event table
POST	/dimension_table	Register a dimension table
POST	/item_table	Register an item table
POST	/scd_table	Register an SCD table
POST	/time_series_table	Register a time series table

Type-specific requirements:

Table Type	Required Fields
Event table	event_timestamp_column
Dimension table	dimension_id_column
Item table	event_id_column, event_table_id
SCD table	effective_timestamp_column, natural_key_column
Time series table	reference_datetime_column, time_interval

All types require a columns_info list describing column names, types, and metadata.

Catalog Tables

Method	Endpoint	Description
GET	/catalog/table	List registered tables in the active catalog
GET	/catalog/table/{table_id}	Get registered table details
GET	/catalog/table/graph	Get table relationship graph

Entities

Method	Endpoint	Description
GET	/entity	List entities
POST	/entity	Create an entity (name + serving_name)

Entity tagging and semantic updates on table columns are typically done via the SDK. See SDK Core Objects for details.

Column Semantics

Apply semantic tags to table columns. Each table type has its own endpoint:

Method	Endpoint	Description
PATCH	/event_table/{id}/column_semantic	Update semantics on an event table
PATCH	/item_table/{id}/column_semantic	Update semantics on an item table
PATCH	/dimension_table/{id}/column_semantic	Update semantics on a dimension table
PATCH	/scd_table/{id}/column_semantic	Update semantics on an SCD table
PATCH	/time_series_table/{id}/column_semantic	Update semantics on a time series table
PATCH	/calendar_table/{id}/column_semantic	Update semantics on a calendar table
PATCH	/snapshots_table/{id}/column_semantic	Update semantics on a snapshots table

Table EDA & Data Cleaning

Method	Endpoint	Description
POST	/table_eda	Run EDA on a registered table (async)
GET	/table_eda	List table EDA runs
PATCH	/table_eda/{id}	Get EDA results
PATCH	/column_analysis	List column analyses (filter by issues)
PATCH	/column_analysis/{id}	Get column analysis details
PATCH	/column_eda/{id}	Get column EDA plots (raw or cleaned)
PATCH	/{table_type}/{table_id}/column_critical_data_info	Set or clear column cleaning operations

EDA parameters: table_id, analysis_size, seed, from_timestamp, to_timestamp, timestamp_column, overwrite.

Contexts

Method	Endpoint	Description
POST	/context	Create a context (primary entities, forecast schema)
GET	/context	List contexts
GET	/context/{context_id}	Get context details

Targets

Method	Endpoint	Description
POST	/target/creation_from_table	Create a target from a table column (async)
GET	/target	List targets
GET	/target/{target_id}	Get target details

Target parameters: target_name, table_id, column_name, offset, fill_value, target_type (REGRESSION, BINARY_CLASSIFICATION, MULTI_CLASS_CLASSIFICATION), positive_label.

Use Cases

Method	Endpoint	Description
POST	/use_case	Create a use case (context + target)
GET	/use_case	List use cases
GET	/use_case/{use_case_id}	Get use case details
PATCH	/use_case/{use_case_id}	Update use case (default EDA table, preview table, name)

Observation Tables

Method	Endpoint	Description
POST	/observation_table/automation	Auto-generate observation table (async)
POST	/observation_table/forecast_automation	Generate forecast observation tables (async)
GET	/catalog/observation_table	List observation tables

Forecast automation parameters: use_case_id, prediction_schedule_cron, prediction_schedule_timezone, forecast_start_offset, forecast_horizon, periods (array with start, end, name, target_observation_count, purpose, mode), dimension_filter (optional).

Modes: ONE_ROW_PER_ENTITY_FORECAST_POINT, FORECAST_SERIES.

Semantic Detection

Automatically detect semantic types for table columns in the context of a use case.

Method	Endpoint	Description
POST	/semantic_detection/column_semantic_detection	Run column semantic detection (async)
GET	/semantic_detection/{id}	Get detection results
GET	/semantic_detection	List semantic detections

Development Dataset

Development Plans

Method	Endpoint	Description
POST	/development_plan/defaults	Get default development plan configuration
POST	/development_plan	Create a development plan (auto-creates dataset)
GET	/development_plan	List development plans (filter by context_id)
GET	/development_plan/{id}	Get development plan details
PATCH	/development_plan/{id}	Update development plan
DELETE	/development_plan/{id}	Delete development plan (async)
GET	/development_plan/{id}/sql_plan	Get sampling SQL plan
PATCH	/development_plan/{id}/sampling_tables	Create distinct entity sampling tables (async)
PATCH	/development_plan/{id}/sampled_tables	Materialize development tables (async)

Development Datasets

Method	Endpoint	Description
GET	/development_dataset	List development datasets
GET	/catalog/development_dataset	List development datasets (filter by use_case_id)
GET	/development_dataset/{id}	Get development dataset details
GET	/development_dataset/{id}/info	Get development dataset info
DELETE	/development_dataset/{id}	Delete a development dataset (async)

Feature Ideation

Generate features automatically based on semantic detection results.

Method	Endpoint	Description
POST	/feature_ideation	Run feature ideation (async)
GET	/feature_ideation	List feature ideations
GET	/feature_ideation/{id}	Get ideation results (optional: include_metadata)

Parameters: use_case_id, semantic_detection_id, windows (aggregation windows).

Ideated Features

Retrieve individual features from an ideation, including SDK code and construction lineage.

Method	Endpoint	Description
GET	/catalog/feature_ideation/{id}/suggested_features	List features with SDK code, relevance scores
GET	/catalog/feature_ideation/suggested_feature_metadata/{id}	Get full SDK code and lineage graph
GET	/catalog/feature_ideation/feature_preview/{id}	Preview computed feature values

Feature EDA

Method	Endpoint	Description
POST	/eda	Run EDA on features (async, optional naive_prediction for residual EDA)
POST	/residual_eda	Run residual EDA on features (async, naive_prediction required)
GET	/eda/{feature_eda_id}	Get feature EDA details (power scores, categories, metadata)
OPTIONS	/eda/{feature_id}/plots	Get available plot options (categories)
GET	/eda/{feature_id}/plots	Get rendered EDA plots for a feature

Input (one of): feature_ids, feature_ideation_id, or feature_list_id.

Feature Selection

SHAP-based feature selection to identify the most impactful features.

Method	Endpoint	Description
POST	/feature_selection	Run SHAP-based feature selection (async)
GET	/feature_selection	List feature selections (supports pipeline_id, use_case_id, feature_ideation_id)
GET	/feature_selection/{id}	Get selection results (scores, rankings)
POST	/feature_selection/{id}/feature_list	Create feature list from selection (async)

Features & Feature Lists

Method	Endpoint	Description
GET	/catalog/feature	List features
GET	/catalog/feature/{feature_id}	Get feature details
GET	/catalog/feature_list	List feature lists
GET	/catalog/feature_list/{feature_list_id}	Get feature list details
POST	/feature_list/job	Create a feature list from feature IDs (async)
POST	/feature_list_from_model	Create feature list from model importance (async)

Feature list from model modes: "Feature importance based", "Feature key importance based".

Model Training

Method	Endpoint	Description
POST	/ml_model	Train a model (async)
GET	/catalog/ml_model	List models (supports pipeline_id, use_case_id, sort_by_metric, show_refits)
GET	/catalog/ml_model/{ml_model_id}	Get model details
POST	/ml_model/{id}/report	Generate model report
OPTIONS	/ml_model/{id}/evaluate	Get evaluation plot options
POST	/ml_model/{id}/evaluate	Create evaluation plot
POST	/ml_model/{id}/refit	Refit model on new data (async)

Model Template Resolution

Method	Endpoint	Description
GET	/use_case/{id}/ml_model_template_setting	Get suggested objective, metric, calibration
GET	/use_case/{id}/ml_model_template	List available model templates

Batch Predictions

Method	Endpoint	Description
POST	/ml_model/{id}/prediction_table	Create prediction table (async)
GET	/catalog/prediction_table	List prediction tables
GET	/prediction_table/parquet/{id}	Download predictions as Parquet
POST	/prediction_table/{id}/prediction_entities	Extract distinct entity values (async)
GET	/prediction_table/{id}/prediction_entities	Get available entity values for forecast comparison
POST	/prediction_table/{id}/forecast_comparison	Create forecast comparison plot (async)
GET	/prediction_table/{id}/forecast_comparison	List forecast comparisons
GET	/prediction_table/{id}/forecast_comparison/{fc_id}	Get forecast comparison plot

Leaderboards

Leaderboards are created automatically when predictions are generated on an observation table with a target.

Method	Endpoint	Description
GET	/catalog/leaderboard	List leaderboards (filter by observation_table_id, observation_table_purpose, role)
GET	/catalog/leaderboard/{leaderboard_id}	Get leaderboard details
POST	/leaderboard/{id}/ml_model/{model_id}/preview	Preview leaderboard predictions

Deployments

Method	Endpoint	Description
POST	/deployment	Create a deployment (async)
GET	/catalog/deployment	List deployments
PATCH	/deployment/{id}	Enable or disable a deployment (async)
POST	/deployment_sql	Generate deployment SQL (async)
GET	/deployment_sql	Get generated deployment SQL

Pipelines (Feature Ideation Orchestration)

Pipelines orchestrate the full ideation workflow as a state machine.

Method	Endpoint	Description
POST	/pipeline	Create a pipeline
GET	/pipeline/{pipeline_id}	Get pipeline status
GET	/pipeline	List pipelines (filter by use_case_id)
PATCH	/pipeline/{pipeline_id}	Advance pipeline (action: advance, optional step_type: end)
PATCH	/pipeline/{pipeline_id}/step_configs	Configure pipeline step
GET	/pipeline/{id}/table_selection/table	List tables available for selection
PATCH	/pipeline/{id}/table_selection	Set selected tables
GET	/pipeline/semantic_detection/column_inputs	Get column inputs from a feature selection (for column pruning)
GET	/pipeline/{id}/feature_ideation	Get suggested features
GET	/pipeline/{id}/feature_selection	Get feature selection results
GET	/pipeline/{id}/validation_leaderboard	Get validation leaderboard
GET	/pipeline/{id}/report	Get pipeline report (JSON)
GET	/pipeline/{id}/report_pdf	Download pipeline report as PDF
GET	/pipeline/{id}/report_md	Download pipeline report as Markdown

Tasks

All async operations return a task ID. Poll for completion:

Method	Endpoint	Description
GET	/task/{task_id}	Get task status and result
GET	/task	List tasks

Task statuses: PENDING, STARTED, SUCCESS, FAILURE, REVOKED.