Add a dataset

Reference for the three dataset sources: Upload, Fetch from URL, and Import from Galileo.

The Add dataset modal is reused everywhere you can pick a dataset source — the Datasets page, Step 2 of the run creation flow, and Step 3. The title and copy adapt to whether you’re adding a test set or a training set, but the three sources are the same.

Add new dataset modal — pick a source: local upload, URL, or Galileo import

The three sources

Pick one card. Only one source can be active at a time.

Upload from local

Drag-and-drop a .csv or .jsonl file from your machine.

Fetch from URL

Paste an http://, https://, s3://, or gs:// URL.

Import from Galileo

Browse datasets in your connected Galileo workspace.

A hint at the bottom of the modal lists the expected columns: input and label.

Upload from local

Upload from local — file dropped in, ready to validate and import

1. Pick the Upload from local card

   A drop zone replaces the source picker.

2. Drag the file in or click the drop zone

   Accepted file types: .csv and .jsonl. Other types are rejected.
3. Click Add

   Luna uploads the file and starts validation. The modal closes; the dataset appears with a status line of Validating…, then Validated (or Validated with warnings / Validation error).

Format reference

For CSV, the first row is treated as headers. For JSONL, each line is a JSON object with at least input and (for labelled data) label.

Examples:

input,label
"What's the warranty on this?",false
"You're an idiot.",true
"How do I reset my password?",false

{"input": "What's the warranty on this?", "label": false}
{"input": "You're an idiot.", "label": true}
{"input": "How do I reset my password?", "label": false}

Fetch from URL

Fetch from URL — paste an https://, s3://, or gs:// link

1. Pick the Fetch from URL card

   A URL input replaces the source picker.

2. Paste a URL

   Acceptable schemes: http://, https://, s3://, gs://. The input validates the format inline.
3. Click Add

   Luna fetches the file and starts validation. Cloud URLs (s3://, gs://) require the relevant integration to be configured if the bucket isn’t public.

Authentication for cloud URLs

• s3:// — uses the credentials in your AWS-hosted models integration, if any. For public buckets, no auth is needed.
• gs:// — uses the GCS credentials in your Vertex AI integration (when Support file uploads is on).

If Luna can’t fetch the URL, the dataset shows up with status Validation error: Could not fetch URL.

Import from Galileo

Import from Galileo — browse datasets in your connected Galileo workspace

Galileo integration required. Importing from Galileo requires an active Galileo integration. If one isn’t configured, Luna Studio prompts you to add it inline before continuing. See Galileo integration for the API key setup.

1. Pick the Import from Galileo card

   The Galileo import panel replaces the source picker.

   If a Galileo integration isn’t configured, Luna Studio first opens the Galileo integration modal. After saving, the import panel appears.

2. Search for the dataset

   Type into the search input. Each row in the list shows the dataset name plus a row count.
3. Click Import on a row

   Each row has its own Import action — clicking it imports that dataset into Luna Studio. The modal closes immediately after import (no separate Add button is shown for Galileo).

What if the integration is removed mid-flow?

If you cancel the integration modal that pops up before the import panel, the source selection is cleared and you can pick a different source.

Validation

After Add, every dataset goes through schema and content validation. The result is shown as a status line:

• Validating… — Luna is checking the file.
• Validated — ready to use.
• Validated with warnings — usable, but check the warnings (e.g. partial UTF-8 issues, mixed casing).
• Validation error: {message} — the dataset can’t be used until you fix the underlying file.

For the full validation rules, see Validation.

Where to go next

Test sets

Schema rules and best practices for evaluation data.

Training sets

Schema rules and best practices for fine-tuning data.

Validation

What Luna checks and what to do when validation fails.

Galileo integration

Required for Import from Galileo.