Create a Datastore

Supported Formats:
- profiles: Parquet, CSV with header
- interactions: JSONLines, CSV
- content: CSV, Parquet
Validation Rules:
- Profiles: must include user_id (string), at least one static feature (e.g. age, region).
- Interactions: require user_id, item_id, timestamp (ISO-8601), optional event_type & value.
- Content: item_id, descriptive features (e.g. title, category, numeric vectors).

Successful Response (201):

Field	Type	Description
datastore_id	string	Unique ID for subsequent calls.
status	string	`ready` (ingestion succeeded), `error`.
ingest_errors	array	Validation failures (if any).

Examples

Get Recommendations

After creating a datastore and training a model, actually get recommendations

Get Recommendations

After creating a datastore and training a model, actually get recommendations

Get Recommendations

After creating a datastore and training a model, actually get recommendations

Check Training Status

Monitor progress and metrics for an active training run

Check Training Status

Monitor progress and metrics for an active training run

Check Training Status

Monitor progress and metrics for an active training run

Start Training a Model

Kick off fine‑tuning on your datastore with a single training_quality setting and optional evaluation dataset.

Start Training a Model

Kick off fine‑tuning on your datastore with a single training_quality setting and optional evaluation dataset.

Start Training a Model

Kick off fine‑tuning on your datastore with a single training_quality setting and optional evaluation dataset.

Create a Datastore

Initializes storage for your client data to train a recommendation model

Create a Datastore

Initializes storage for your client data to train a recommendation model

Create a Datastore

Initializes storage for your client data to train a recommendation model

Go Back