DUX “Paid invoices” export to Google Sheets (no DUX API)
Budget: $30 – $250 USD
Project description
I need a reliable way to detect which sales invoices/remitos are marked as paid in the DUX web UI (erp.duxsoftware.com.ar) and write them into a Google Sheet.
What I want
* Source of truth is the DUX “Facturación → Ventas” list screen where each row shows at least: Invoice/Comprobante (e.g., “CX-00001-00020623”), Total, and Cobrado.
* A row is considered PAID when: Cobrado == Total AND Cobrado > 0 (must parse AR currency formats: “$ 1.234,56” etc.).
* For each PAID row, extract the invoice number in format “00001-00020609” (in DUX it appears as “CX-00001-00020609”; extract with regex \d{5}-\d{8}).
* Append results into this Google Sheet (tab: “Cobrados”), columns:
1. Fecha (timestamp when detected)
2. Remito (00001-00020609)
3. Cobrado (amount)
Google Sheet link: [https://docs.google.com/spreadsheets/d/1tJYnTkrqMJCbmCeBYgK-PxMPqrK0KjNkBN7iJXt8fzM/edit?gid=0#gid=0](https://docs.google.com/spreadsheets/d/1tJYnTkrqMJCbmCeBYgK-PxMPqrK0KjNkBN7iJXt8fzM/edit?gid=0#gid=0)
* Must avoid duplicates (if the Remito already exists in the sheet, don’t add it again).
Key constraint
* DO NOT use DUX payments API endpoints. We tried API-based approaches and they are unreliable / rate-limited for this use case. The solution must rely on what the DUX web UI shows (any approach you prefer: browser automation, DOM extraction, network inspection, etc.).
How often / how it runs
* I’m open to your recommendation: manual (button/script run), scheduled (every 15 minutes), or both. The only requirement is that it’s practical for a Windows PC user and stable.
What we already tried (so you don’t repeat dead-ends)
1. Network “listener” approach:
* We attempted to detect paid rows by intercepting browser traffic while a user navigates the DUX Ventas page (capturing HTML / JSF partial responses / XHR responses) and parsing Total/Cobrado from those responses.
* In practice, we couldn’t make it stable enough across sessions and UI changes.
2. Separate automation/scraper approach:
* We considered a standalone script that logs into DUX, navigates to Facturación → Ventas, paginates through the last N pages, reads the visible table values (Total/Cobrado/Comprobante) and exports the paid ones.
* We didn’t finalize a working implementation.
Deliverable
* A working solution with clear setup instructions and a simple way to run it on Windows.
* Logging/reporting: number of pages scanned, number of paid invoices found, duplicates skipped, and any errors.
What I will provide after hiring
* A DUX user account for testing (shared privately).
* The Google Sheet access as needed.
* Screenshots of the relevant DUX screen (Facturación → Ventas) showing the columns Total and Cobrado.
What I will NOT provide
* My previous source code attempts (interceptor/launcher scripts) — you can implement it your own way.
I need a reliable way to detect which sales invoices/remitos are marked as paid in the DUX web UI (erp.duxsoftware.com.ar) and write them into a Google Sheet.
What I want
* Source of truth is the DUX “Facturación → Ventas” list screen where each row shows at least: Invoice/Comprobante (e.g., “CX-00001-00020623”), Total, and Cobrado.
* A row is considered PAID when: Cobrado == Total AND Cobrado > 0 (must parse AR currency formats: “$ 1.234,56” etc.).
* For each PAID row, extract the invoice number in format “00001-00020609” (in DUX it appears as “CX-00001-00020609”; extract with regex \d{5}-\d{8}).
* Append results into this Google Sheet (tab: “Cobrados”), columns:
1. Fecha (timestamp when detected)
2. Remito (00001-00020609)
3. Cobrado (amount)
Google Sheet link: [https://docs.google.com/spreadsheets/d/1tJYnTkrqMJCbmCeBYgK-PxMPqrK0KjNkBN7iJXt8fzM/edit?gid=0#gid=0](https://docs.google.com/spreadsheets/d/1tJYnTkrqMJCbmCeBYgK-PxMPqrK0KjNkBN7iJXt8fzM/edit?gid=0#gid=0)
* Must avoid duplicates (if the Remito already exists in the sheet, don’t add it again).
Key constraint
* DO NOT use DUX payments API endpoints. We tried API-based approaches and they are unreliable / rate-limited for this use case. The solution must rely on what the DUX web UI shows (any approach you prefer: browser automation, DOM extraction, network inspection, etc.).
How often / how it runs
* I’m open to your recommendation: manual (button/script run), scheduled (every 15 minutes), or both. The only requirement is that it’s practical for a Windows PC user and stable.
What we already tried (so you don’t repeat dead-ends)
1. Network “listener” approach:
* We attempted to detect paid rows by intercepting browser traffic while a user navigates the DUX Ventas page (capturing HTML / JSF partial responses / XHR responses) and parsing Total/Cobrado from those responses.
* In practice, we couldn’t make it stable enough across sessions and UI changes.
2. Separate automation/scraper approach:
* We considered a standalone script that logs into DUX, navigates to Facturación → Ventas, paginates through the last N pages, reads the visible table values (Total/Cobrado/Comprobante) and exports the paid ones.
* We didn’t finalize a working implementation.
Deliverable
* A working solution with clear setup instructions and a simple way to run it on Windows.
* Logging/reporting: number of pages scanned, number of paid invoices found, duplicates skipped, and any errors.
What I will provide after hiring
* A DUX user account for testing (shared privately).
* The Google Sheet access as needed.
* Screenshots of the relevant DUX screen (Facturación → Ventas) showing the columns Total and Cobrado.
What I will NOT provide
* My previous source code attempts (interceptor/launcher scripts) — you can implement it your own way.
Related categories:
JavaScript
Python
Web Scraping
Node.js
Scripting
Web Development
Data Extraction
Google Sheets
Automation