Step-By-Step Instruction

Follow these steps to integrate and use the Agent Document Parser in your OutSystems

application:

How to Install, Use, and Configure

Installation

1. Prepare Your Document

Gather the document you want to parse. Supported formats include:

PDF

Image (e.g., JPG, PNG)

Plain text

2. Define Your Output Schema

Create a JSON schema that describes the structure of the data you want to extract.

Example:

This schema tells the agent exactly which fields to look for and how to structure the output.

{

"name": "string",

"email": "string",

"phone": "string",

"yearsOfExperience": "number"

}

3. Upload the Document in Your Application

Use the UploadFileInfo component to retrieve the file type of the document uploaded in

your app.

4. Call the Agent Document Parser Action

Pass the uploaded file (as binary), its file type (e.g., "application/pdf"), and your JSON

schema to the agent’s parsing action.

5. Receive and Use Structured JSON Output

The agent returns a JSON object that strictly follows your schema, filled with data extracted

from the document. You can now use this structured data in your workflows, databases, or

UI.

1. Go to the OutSystems Forge .

2. Search for “Agent Document Parser”.

Usage

Once installed, the component exposes a service action CallAgentDocumentParser

Input Parameters:

Output:

Configuration Tips

Example Use Case: Resume Parsing

3. Click “Install” to add the component to your environment.

FileContent (Binary Data): The raw content of the uploaded file.

FileType (Text): MIME type of the file (e.g., "application/pdf" , "image/jpeg" ,

"text/plain" ).

→ Tip: Use the UploadFileInfo Forge component to auto-detect this.

JsonSchema (Text): A valid JSON object defining the expected output structure.

ParsedResult (Text): A JSON string that matches your schema, populated with extracted

data.

Success (Boolean): Indicates whether parsing succeeded.

ErrorMessage (Text, optional): Contains error details if parsing failed.

File Type Accuracy: Ensure the FileType parameter matches the actual file format.

Incorrect types may lead to parsing errors.

Schema Design: Keep your JSON schema simple and aligned with the document’s

content. The agent works best when field names reflect common labels (e.g., “Email”,

“Name”, “Date of Birth”).

Error Handling: Always check the Success flag and handle errors gracefully in your

application logic.

Testing: Start with small, clean documents (e.g., a one-page resume) to validate your

schema before scaling to complex forms.

1. User uploads a CV (PDF).

2. Your app sends the file + schema (with fields like name , email , skills ) to the agent.

3. Agent returns structured JSON.

4. Your app auto-fills a candidate profile in the database.