Web client

The web dashboard lets you manage projects and document types without writing code. Log in at app.helvetii.ai to get started.

Creating a project

Projects are the top-level container for organizing your document types. Each project has its own set of document types, schemas, and extraction routines.

  1. 1

    Log in to the dashboard

    Navigate to app.helvetii.ai and sign in with your account credentials. Authentication is handled through Auth0.

  2. 2

    Navigate to Projects

    Click Projects in the sidebar navigation. You'll see a list of your existing projects, or an empty state if this is your first time.

    Projects list view
  3. 3

    Click "New project"

    A modal dialog opens with the project creation form.

  4. 4

    Fill in project details

    Enter the following fields:

    nameAlphanumeric characters, hyphens, underscores, and dots. No spaces.
    descriptionA brief description of the project's purpose.

    Note

    The project name is used as an identifier in API calls. Choose a descriptive, URL-safe name like insurance-claims or invoice-processing.

    Create project modal
  5. 5

    Submit

    Click the submit button. The project is created and appears in your projects list. You can now add document types to it.

Creating a document type

Document types define the extraction schema for a category of documents within a project. Each document type specifies which fields to extract and how to validate the results.

  1. 1

    Navigate to Document Types

    Go to Document Types in the sidebar, or click Add document type from a project detail page.

  2. 2

    Click "New document type"

    The creation modal opens. If you navigated from a project, the project field is pre-selected.

  3. 3

    Enter name and description

    The document type name follows the same rules as project names: alphanumeric, hyphens, underscores, dots. No spaces. Examples: invoice, electricity-bill, claim-form.

  4. 4

    Define the extraction schema

    The schema defines which fields the AI should extract from documents. Each field has a name, type, and description. The description guides the model on what to look for.

    Example: invoice schema
    {
  "invoice_number": {
    "type": "string",
    "description": "Unique invoice identifier"
  },
  "date": {
    "type": "string",
    "description": "Invoice date in YYYY-MM-DD format"
  },
  "vendor": {
    "type": "string",
    "description": "Vendor or supplier name"
  },
  "total_amount": {
    "type": "number",
    "description": "Total amount including tax"
  },
  "currency": {
    "type": "string",
    "description": "ISO 4217 currency code"
  },
  "line_items": {
    "type": "array",
    "description": "List of invoice line items",
    "items": {
      "description": { "type": "string" },
      "quantity": { "type": "number" },
      "unit_price": { "type": "number" }
    }
  }
}

    See schema best practices for guidance on writing descriptions that extract reliably.

  5. 5

    Configure advanced options

    Expand the Advanced options section to configure:

    validation_rulesJSON array of business rules applied after extraction.
    field_importanceJSON object mapping field names to importance levels: critical, high, medium, low.
  6. 6

    Submit

    The document type is created and linked to your project. You can now process documents against it using the API.

    Create document type modal