解析文档 n8n

Q: How does the n8n node know what fields to extract?

The capture keys are defined in the dashboard, not in n8n. Each capture area in the template gets a key name (camelCase) and an extraction rule: Regex Expression for stable patterns (invoice numbers, dates, amounts) or JavaScript Expression for conditional logic. The output JSON has one field per key.

Q: Can I extract data using JavaScript Expression in addition to regex?

Yes. The template engine exposes the extracted text as the variable text inside your JavaScript Expression and you return a string. Use it for multi-marker classification (invoice vs order vs receipt), fallback rules, or date format normalization. See the Prepare Parse Info for Document guide for two working samples.

此节点的功能

PDF4me 解析文档 使用您保存的解析模板对 PDF 在 n8n 工作流并将提取的字段作为结构化数据返回。向其提供来自先前节点的二进制数据。 Base64 字符串，或公共字符串 URL引用模板 解析 ID并将结果通过管道传输。 JSON， XML，或者 CSV 直接进入 Set、IF、HTTP 请求或任何下游 n8n 节点。模板包含提取逻辑，因此同一个节点可以提取发票、合同、收据以及您配置的任何自定义布局。

验证您的身份 API 要求

这 PDF4me 节点 n8n 需要 PDF4me API 证书只需创建一次凭据即可。 n8n 你的 API 从仪表板获取密钥，然后从每个地方引用它。 PDF4me 工作流程中的节点。

Get your API Key

您不容错过的重要事实

解析 ID 为必填项（即您的 TemplateId）。

解析 ID 是仪表板在模板上保存更改时分配的 GUID。从模板详细信息面板复制它，然后将其粘贴到 n8n 节点。如果没有它，节点就没有要提取哪些字段的映射。

用于稳定模式的正则表达式 JavaScript 用于条件逻辑

捕获键存储在模板中。发票编号、日期、金额、税号（约占生产键的 80%）使用正则表达式。 JavaScript 用……表达 文本 用于分类和回退规则的变量。

输出是结构化数据，而不是二进制数据。 PDF

与返回原始数据的压缩、保护或转换节点不同。 PDF 字节，解析文档返回 JSON （默认）， XML，或者 CSV 在指定的二进制字段中。将其连接到 Set 或 IF 节点，以便将值路由到下游。

参数

范围	必需的	它的作用	例子
Input Data Type	Yes	How the PDF reaches the node. Binary Data (from a prior node), Base64 String (inline encoded), or URL (public file URL).	`Binary Data`
Input Binary Field	Conditional	Name of the binary field on the input item containing the PDF. Required when Input Data Type is Binary Data.	`data`
Base64 PDF Content	Conditional	Base64 encoded PDF content. Required when Input Data Type is Base64 String. No data: prefix.	`JVBERi0xLjQK...`
PDF URL	Conditional	Publicly reachable URL of the PDF. Required when Input Data Type is URL. The PDF4me service downloads and processes it.	`https://example.com/invoice.pdf`
Document Name	Yes	Source filename including .pdf extension. Used for format detection and error tracing.	`invoice.pdf`
Parse ID	Yes	TemplateId GUID from the PDF4me dashboard. Identifies the parse template (capture keys + extraction rules) to apply.	`12345678-1234-1234-1234-123456789abc`
Output Format	Yes	Shape of the returned data. JSON (default, structured object), XML (hierarchical), or CSV (flat tabular). JSON is the right choice for almost every n8n workflow.	`JSON`
Output Binary Field Name	Yes	Name of the binary field the node attaches the parsed output to on the output item. Reference it from downstream nodes.	`data`
Custom Profiles	No	Advanced JSON-style overrides for parsing behaviour (output sub-format, include metadata flag, etc). Leave empty for default behaviour.	`{ "outputDataFormat": "json", "includeMetadata": true }`

解析模板中的表达式类型

捕获规则存在于模板中，而不是在……中。 n8n 节点。两种表达式类型都随每个节点一起提供。 PDF4me 帐户。

正则表达式稳定模式

发票号码（INV-\d{6,10}），日期（\d{2}/\d{2}/\d{4}），金额（$?\d{1,3}(?:,\d{3})*(?:.\d{2})?约占生产密钥的 80%。

JavaScript 表达式条件逻辑和分类器

捕获区域文本以如下方式传递： 文本返回一个字符串。支持多标记分类、回退规则和日期规范化。参见准备文档解析信息提供两个工作样品。

输出

场地	类型	它包含什么
`fileName`	String	Generated output filename with timestamp and extension matching Output Format (.json / .xml / .csv).
`mimeType`	String	MIME type of the output (application/json, application/xml, text/csv).
`fileSize`	Number	Size of the parsed output in bytes.
`success`	Boolean	true on a successful parse, false otherwise. Wire into an IF node for branching.
`message`	String	Status message. Document parsed successfully on the happy path, or a descriptive error.
`docName`	String	Original input filename, preserved for audit trails.

解析后的键/值数据本身存储在您指定的二进制字段中。 输出二进制字段名称。

工作流程示例

常见的 n8n 工作流程模式Typical ways to chain Parse Document into an n8n workflow.

通过电子邮件发送到 Postgres。

Gmail 触发器会在新发票标记为 vendor-invoice 时触发。
筛选器仅通过以 .pdf 结尾的附件。
解析文档会针对二进制附件运行发票模板。
已解析集合节点映射 invoiceNumber， totalAmount， invoiceDate。
Postgres Insert 语句会将一行写入 accounts_payable 表中。

在一个工作流程中完成分类和提取

Webhook 收到一个 PDF 来自合作伙伴的上传门户。
解析文档会运行一个带有模板的程序 JavaScript 返回文档类型的表达式键。
根据返回的类型（发票/订单/收据）切换节点路由到特定类型的下游分支。
每个分支都会将解析后的字段发布到正确的目标系统。

将 S3 中的数据批量解析到 Airtable

定时触发器每 15 分钟触发一次。
S3 List + Get Object 加载每个新对象 PDF 在入库桶中。
解析文档会针对每个二进制项运行模板。
Airtable 创建记录会将解析后的字段写入跟踪数据库。

常见问题解答

What is a Parse ID and where do I get it?+

Parse ID is the TemplateId GUID generated by the PDF4me dashboard when you click Save Changes on a parse template. Find it in the template detail panel. Pin it in your n8n node so the same template runs every execution.

Do I need a template, or can the node parse any PDF without one?+

A template is required for field-level extraction in n8n. Without a Parse ID the node has no map of which regions to capture. Build a template once in the dashboard, then reuse the same Parse ID across runs and across platforms (Make, Zapier, Power Automate).

How does the n8n node know what fields to extract?+

The capture keys live in the template, not in n8n. Each capture area gets a key name (camelCase) and an extraction rule: Regex Expression for stable patterns or JavaScript Expression for conditional logic. The output has one field per key.

Which Input Data Type should I pick?+

Binary Data for files arriving from upstream nodes (Read Binary Files, HTTP Request with response format File, Gmail attachment, Dropbox Download). Base64 String when the PDF is encoded inline. URL when the file lives at a public web address.

Can I extract data using JavaScript Expression in addition to regex?+

Yes. The capture area text is passed as the variable text inside your JavaScript Expression and you return a string. Use it for multi-marker classification, fallback rules, or date normalization. See the Prepare Parse Info for Document guide for two working classifier samples.

JSON, XML, or CSV. Which Output Format should I use?+

JSON for almost every n8n workflow. It maps cleanly into Set, IF, and HTTP Request nodes. XML when your downstream system requires it natively. CSV for spreadsheet uploads and bulk imports.

Is the output a file I can save, or just a JSON object?+

Both. The parsed data is attached to the binary field you set under Output Binary Field Name, so you can Write Binary File to disk, send it as an email attachment, or upload it to storage. The structured data is also available on the regular json output for downstream node mapping.