> ## Documentation Index
> Fetch the complete documentation index at: https://docs.recordengine.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Integrations Overview
> How RecordEngine connects to the rest of your software stack — without sending any data to the cloud.
RecordEngine is built integration-first. Every document, every extracted field, and every status change is accessible via REST API and outbound webhooks — so the structured data the AI extracts can flow directly into Salesforce, QuickBooks, Xero, or any other system your team uses.
All of this happens **without any document data leaving your server**.
***
## Three Integration Patterns
All RecordEngine integrations are built from a combination of these three patterns:
Your external system (CRM, ERP, email) sends a document to RecordEngine via the API, including a reference ID that links the document back to the originating record.
RecordEngine fires an outbound webhook when a document is approved or exported. Your external system receives the full structured payload and updates its records automatically.
A URL stored in your CRM opens the exact document in RecordEngine directly — so your team can click from a Salesforce opportunity straight to the supporting invoice, without searching.
Most integrations use all three patterns together to create a fully bidirectional pipeline.
***
## The CRM Correlation Key
The glue that makes bidirectional integrations work is a field called **`external_refs`**. When a document is uploaded via the API, you can attach a reference to the originating CRM record:
```json theme={null}
{
"external_refs": {
"salesforce": {
"record_id": "001Qy00000BnXt2IAF",
"record_type": "Opportunity",
"record_url": "https://yourorg.salesforce.com/001Qy00000BnXt2IAF"
}
}
}
```
This reference is:
* **Stored** on the document in RecordEngine
* **Displayed** as a clickable link in the document detail panel (🔗 External Links)
* **Echoed** in every outbound webhook payload — so when the document is approved, your automation platform knows exactly which Salesforce record to update
See [How CRM Correlation Works](/integrations/how-crm-correlation-works) for the full technical details.
***
## Available Integrations
Two-way sync with Opportunities, Contacts, and Accounts. Store extracted invoice data on Salesforce records. Deep-link from Salesforce straight to the document.
Attach extracted data to HubSpot Deals and Contacts. Trigger HubSpot workflows when documents are approved.
Bidirectional sync with Dynamics 365 records via the same CRM correlation pattern as Salesforce.
Create Bills and Invoices in QuickBooks automatically from RecordEngine's extracted line items and totals.
Push approved invoices and receipts to Xero as draft Bills — ready for your accountant to review.
Send document approval notifications to WeCom (Enterprise WeChat) group channels.
Connect RecordEngine to any system using no-code automation platforms and the webhook + API layer.
Build your own integration using the outbound webhook payload. Full schema, code examples in Python and Node.js.
***
## How the Data Flows
Here is a complete end-to-end example for a Salesforce integration:
```
1. Salesforce Opportunity is created
↓
2. Automation triggers → calls RecordEngine API to upload the supporting document
(includes external_refs.salesforce.record_id)
↓
3. RecordEngine AI extracts fields, scores confidence
↓
4. Reviewer approves document in RecordEngine
↓
5. RecordEngine fires outbound webhook with full payload
(includes extracted fields + external_refs.salesforce.record_id)
↓
6. Automation receives webhook → reads record_id
→ updates Salesforce Opportunity with vendor, amount, invoice number
↓
7. Deep-link URL stored on Salesforce record
→ Sales rep clicks link → jumps directly to document in RecordEngine
```
***
## No Data Leaves Your Server
Every integration uses **outbound webhooks and inbound API calls** — RecordEngine never initiates a connection to Salesforce, QuickBooks, or any other cloud service directly. Your automation platform acts as the intermediary, running on its own infrastructure.
This means:
* RecordEngine operates fully offline — it has no cloud dependencies
* Document content is never transmitted to a third-party AI or cloud service
* You control exactly which fields are shared with which external systems
***
## Getting Started
The fastest way to start integrating is:
Go to **Settings → Webhooks** and enter the URL that should receive document payloads when documents are exported.
Go to **Settings → API** and copy your Bearer token. This is used to authenticate all inbound API calls.
Choose the platform you want to connect from the cards above and follow the step-by-step guide.
If you need help designing a custom integration, contact us at [hello@recordengine.ai](mailto:hello@recordengine.ai).