Datasets

Onboard a dataset — zero-config schema profiling + auto-training

Drop in data → instantly queryable.

  • Accepts JSON records, inline CSV, or column names only
  • Profiles schema and role-classifies every column automatically
  • Auto-suggests domain aliases (PPA → per/person/average, etc.)
  • Fires background semantic training immediately when the training runtime is installed
  • Returns model tier — queries work immediately via base/deterministic fallback

For file uploads (CSV, Excel, Parquet, JSON) use POST /v1/datasets/onboard/file.

post/v1/datasets/onboard

Request

  • Base URL: https://api.algenta.ai
  • URL: https://api.algenta.ai/v1/datasets/onboard
  • Auth: one of:
    • HTTP bearer
    • API key in header X-API-Key

Request body

namestring

Human-readable dataset name

source_namestring nullable

Table name used in the schema

recordsobject[] nullable

JSON records (list of row dicts)

csvstring nullable

Raw CSV text with header row

columnsstring[] nullable

Column names only — no data required

domain_aliasesobject nullable

Manual alias map. Auto-suggested from column names when omitted.

async_trainboolean

Fire background semantic training immediately

epochsinteger
batch_sizeinteger

Response

Successful Response

dataset_idstring required
namestring required
schema_hashstring required
statusstring required
model_tierstring required
decision_modestring required
source_namesstring[] required
column_countinteger required
domain_aliasesobject nullable required
registered_atstring nullable required
updated_atstring nullable required
schemaobject required
suggested_aliasesobject required
latency_msnumber required

Changes

No recorded changes to this operation across all 1 revision of this API.