AI Agent Hub
Back to skills
Xiangyun Invoice OCR icon

Xiangyun Invoice OCR

Office Efficiency Updated 2026.08.30

Paste the following prompt into your AI chat to install this skill:

Please follow https://skillhub.cn/install/skillhub.md to install @user_6b190ef3/invoice-ocr-xy.

About this skill

Problem

Invoice data often sits in email threads, chat logs, and drive folders, while manual review of invoice type, format, and readability is easy to miss. This skill turns PDF, OFD, and common image invoices into a callable recognition workflow for extracting basic invoice fields in batch engineering scenarios.

How It Works

The skill relies on Xiangyun OCR and uses netocr_key and netocr_secret for authentication. After input files are provided, it requests recognition through the Xiangyun API and parses invoice type and key fields, covering 40+ invoice types such as VAT special invoices, digital invoices, train tickets, toll invoices, and bank receipts. Typical steps include:
- Prepare readable files, each under 10MB
- Support .pdf, .ofd, .jpg, .png, and .bmp
- Call the recognition API and parse the response
- Feed the structured result into archiving, reconciliation, or reimbursement flows

Boundaries

Accuracy depends on image clarity, file format, and external API quota. Images should be at least 500px on the relevant side; small or blurry files may miss fields. Xiangyun OCR is billed per call, so estimate file volume and cost before running large batches. The config file is stored as config.json under the skill directory, and credentials should not be committed to repositories.

Use Cases

  • Finance staff receives many PDF and OFD e-invoices at month end and needs type and key fields for reimbursement.
  • A dev team batch-processes train, toll, and ride receipts from a download folder into structured records.
  • Shared services checks medical receipts, bank receipts, and VAT invoices before routing them to manual review.
  • An expense platform calls this skill after submission to recognize PDF, OFD, or image invoices and validate readability.

Best For

  • Finance staff handling monthly reimbursement: needs scattered PDF, OFD, and image receipts turned into structured data.
  • Engineers building expense platforms: needs invoice OCR in the submission flow and parsed invoice-type fields.
  • Shared services agents: needs to distinguish VAT, train, toll, and medical receipts before manual review.
  • Admin staff archiving receipts: needs batch recognition of common invoice types and readable file metadata.