跳到正文
TakoAPI
返回技能列表

documents-ai

由 dbirulia 提供

Speech & Transcription
559下载量由 clawskills.sh 提供
16星标
0浏览量
0 个赞登录后即可为这个技能点赞
—暂无评分
你的评分

简介

Real-time OCR and data extraction API by Veryfi.

描述

该技能的功能

OCR and structured data extraction API by Veryfi. Processes receipts, invoices, bank statements, checks, W-9s, passports, and dozens of other document types. Returns parsed fields, line items, fraud scores, and raw OCR text.

Veryfi returns structured JSON with categorized line items, vendor details, and fraud scores in a single API call, eliminating the need to build and maintain custom OCR parsing logic.

何时使用

  • Extract line items from a vendor invoice for accounts payable
  • Parse bank statements to categorize monthly transactions
  • Scan receipts for expense report automation
  • Classify incoming documents before routing to processing pipelines
  • Pull structured fields from W-9 forms during contractor onboarding

示例工作流

输入

User asks: extract all transactions from this bank statement PDF

智能体
  1. 1

    Reads the PDF file and encodes it as base64 or prepares it for multipart upload

  2. 2

    Posts the file to /api/v8/partner/bank-statements/ with VERYFI_CLIENT_ID and VERYFI_API_KEY headers

  3. 3

    Receives JSON response containing account details, beginning/ending balances, and a transactions array

  4. 4

    Extracts the transactions array with date, description, credit_amount, debit_amount, and category fields

  5. 5

    Returns a formatted summary of all transactions to the user

输出

Structured list of transactions with dates, amounts, and spending categories from the statement period

前置要求

  • VERYFI_CLIENT_ID environment variable
  • VERYFI_USERNAME environment variable
  • VERYFI_API_KEY environment variable
  • Veryfi account (sign up at https://app.veryfi.com)

安装

clawhub install dbirulia/documents-ai

运行 CLI 命令,或把这段提示词粘贴到 Claude Code、Cursor、Codex 或你在用的任意编程智能体里。

链接