Add Data overview

Use Datasets > Add Data to bring source material into RebelCore. You can select or drag several files at once; RebelCore classifies each item and sends the resulting module down the appropriate path.

Add files → Review classification → Create module → Create project

The steps after module creation depend on what the files can do:

Input	RebelCore creates	Where you continue
Excel or CSV only	A structured module	Module Builder, then the project Tree
Documents only	A vectorised file module	Project creation, then RebelCore Agent
Tabular data and documents	A mixed module	Module Builder and Tree, with documents available to Agent

Before you start

Confirm that:

• Your role allows you to view and add datasets.
• The material belongs to the organisation shown in RebelCore.
• You are authorised to upload and use the information it contains.
• Related files belong in one module and should share one LLM exposure policy.
• Tabular files have clear headings and consistent structures.

Remove unnecessary personal or confidential information where your organisation’s handling policy requires it.

1. Add and classify the files

1. Open Datasets under Workspace.
2. Select Add Data.
3. Enter a meaningful module name.
4. Drag files onto the blade or choose them from your device.
5. Review the groups RebelCore creates:
   • Tabular data
   • Source files
   • Needs review
6. Resolve every item under Needs review.
7. Choose the LLM exposure level that should govern the resulting module.

RebelCore keeps tabular data and source documents visibly separate. Classification controls the workflow; it does not weaken file validation or security scanning.

Add files

2. Follow the recommended path

Tabular-only

The upload opens Module Builder. Review files and worksheets, configure the schema, use AI Data Advisor where useful, confirm the exposure level, and build the module.

Documents-only

RebelCore stores, parses, chunks, and schedules the documents for vectorisation. It then opens Create Project with the file module selected. The resulting project opens directly in Agent because there is no tabular Tree to train.

Mixed

The tabular portion opens Module Builder. A read-only Source Files tab lists the documents carried with the module and their vectorisation status. The exposure selected for the module governs both its tabular and file capabilities.

Create your dataset

3. Create and open the project

Create a project from one or more ready modules and choose who may access it.

• A project with only file modules opens directly in Agent.
• A project with tabular or mixed modules opens in Tree for curation and semantic training.
• Dataset-associated documents become available to users through project access; they are not copied into personal file shares.

Creating a project

Understand status

Status	Meaning
Processing	Tabular import work is still running.
Created	A tabular or mixed source is ready for Module Builder.
Vectorising	File content is being parsed and embedded.
Partial	Some source documents are usable and others require attention.
Complete	The module is ready for its next step.
Failed	Review the item-level error and retry or replace the affected source.

Refresh the Datasets page when waiting for background work. Opening a dataset row shows tabular files and source documents in separate groups.

Next steps

• Add files
• Create your dataset
• LLM exposure levels
• Creating a project