Skip to content

UiPath connector

Updated

Overview#

DocXtract Activities is a UiPath package with one activity, Extract Document: point it at a file and it returns the structured JSON. No HTTP Request activity, no multipart plumbing. It calls the same v3.1 API as every other integration, so accuracy, credit cost, and document limits are identical.

Before you start#

  • An API key from the client portal → Settings
  • The document types your key permits — check on the Test API page
  • Documents as PDF, JPG, or PNG, up to 10 MB
  • UiPath Studio 22.10 or later, on a Windows or Windows-Legacy project
Tip
Give this connector its own API key. Usage stays attributable per tool, and you can revoke it without disrupting your other integrations.

Setup#

1. Install the package#

In Studio open Manage Packages, search for DocXtract Activities, and install it into your project. If the package is not in your organisation's feeds, ask support@docxtract.io for the .nupkg and add it through a local or Orchestrator feed.

Screenshot pending

Studio → Manage Packages showing DocXtract Activities and the Install button

2. Add the activity#

Drag Extract Document from the Activities panel into your sequence, after whatever step obtains the file — a download, an email attachment save, or a folder watch.

Screenshot pending

Extract Document activity placed in a workflow sequence

3. Configure it#

PropertyRequiredValue
File PathYesFull path to the document on the robot machine
API KeyYesYour DocXtract key — read it from an Orchestrator credential asset, never a hard-coded string
API URLYeshttps://api.docxtract.io/v3.1/documents
ModelYesLeave as default unless support has given you a specific model name
Document TypeYesinvoice, purchase_order, or any other type your key permits
JSON ResponseOutputA string variable that receives the result
Screenshot pending

Properties panel showing File Path, API Key, API URL, Model, Document Type and the JSON Response output

4. Read the result#

JSON Response holds the API envelope. Deserialize it and read data:

{
  "success": true,
  "data": {
    "vendor": "ABC Pvt Ltd",
    "invoice_no": "INV-1001",
    "invoice_date": "2024-05-20",
    "currency": "INR",
    "total": 12500
  },
  "processing_time_ms": 3420,
  "model_used": "default"
}

Check success before touching data, and branch on error.code for failures — the message text is for humans and may change.

What to expect#

  • Credits are per page. A 10-page document costs 10 credits.
  • Rate limits are per key — 10 requests per minute by default. An unattended robot looping over a folder will hit this; add a short delay or a retry scope with backoff.
  • PDFs over 3 pages are split and processed in parts, so they take noticeably longer than single-page documents. Allow for that in the activity's timeout.
  • data is an open map that varies by document type. Invoice and purchase order are the types validated for this release; others work if they are on your key.

Troubleshooting#

SymptomCause
Authentication failsKey is revoked, expired, or the asset value is not a DocXtract sk_ key
400 unknown_modelThe document type is not assigned to this key — check the portal
Extraction returns unexpected fieldsWrong document type; it defaults to invoice
Requests start failing during a bulk runRate limit — 10 requests/minute per key
Activity not found after installProject targets a framework the package does not support — use Windows or Windows-Legacy
Stops working suddenly, mid-monthCredit cap reached; see quotas

More detail in common problems.

Support#

support@docxtract.io — include the document type and roughly when the failure happened. Never send your API key.

Related