Setup Nanonets (OCR + Review with Learning) + Webhook for Line-Level JSON (CUG)
Budget: €8 – €1,000 EUR
I’m looking for a developer to configure Nanonets (or an equivalent low-cost IDP) to read invoice line items, and to build a simple webhook that receives the approved output and maps it to a CUG (line-level JSON payload) ready for my system.
I need the setup to learn from my corrections in the Nanonets Review UI, and I want a brief training session on how to handle new suppliers in the future.
Scope (ONLY 2 tasks)
1. Nanonets
• Create an invoice line-item model with core fields (supplier/customer, product code/description, lot, expiry, qty, UOM, unit price, line total, discount, VAT, invoice number/date).
• Enable the Review UI and learning from corrections.
• Configure export via webhook for approved documents/lines only.
2. Webhook (CUG)
• Implement a small service (FastAPI or Node) that receives Nanonets JSON and maps it into a CUG (line-level JSON format we’ll define together).
• Include basic dedup (hash), light string normalization, and minimal logging.
• Deliver code + README (Docker welcome).
Deliverables
• Access to the configured Nanonets model.
• Source code for the webhook + deployment instructions.
• Handover session (screen share) to show:
• how to correct in Review so the model “learns,”
• how to onboard/manage new, never-seen suppliers,
• how to adjust the CUG mapping if fields change.
Requirements
• Experience with Nanonets (or similar IDP) and webhooks.
• Ability to provide a short training at the end.
Budget
• Under $1,000 total for these two tasks.
How to apply (brief)
• 1 link or short description of a similar project (OCR → review → webhook).
• Your preferred stack for the webhook (FastAPI or Node).
• Confirm you can include the short training session.
I need the setup to learn from my corrections in the Nanonets Review UI, and I want a brief training session on how to handle new suppliers in the future.
Scope (ONLY 2 tasks)
1. Nanonets
• Create an invoice line-item model with core fields (supplier/customer, product code/description, lot, expiry, qty, UOM, unit price, line total, discount, VAT, invoice number/date).
• Enable the Review UI and learning from corrections.
• Configure export via webhook for approved documents/lines only.
2. Webhook (CUG)
• Implement a small service (FastAPI or Node) that receives Nanonets JSON and maps it into a CUG (line-level JSON format we’ll define together).
• Include basic dedup (hash), light string normalization, and minimal logging.
• Deliver code + README (Docker welcome).
Deliverables
• Access to the configured Nanonets model.
• Source code for the webhook + deployment instructions.
• Handover session (screen share) to show:
• how to correct in Review so the model “learns,”
• how to onboard/manage new, never-seen suppliers,
• how to adjust the CUG mapping if fields change.
Requirements
• Experience with Nanonets (or similar IDP) and webhooks.
• Ability to provide a short training at the end.
Budget
• Under $1,000 total for these two tasks.
How to apply (brief)
• 1 link or short description of a similar project (OCR → review → webhook).
• Your preferred stack for the webhook (FastAPI or Node).
• Confirm you can include the short training session.