> ## 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.

# Lark / Feishu Integration

> Automatically ingest documents into RecordEngine by dropping files into a Lark or Feishu group chat.

## Overview

RecordEngine can receive documents directly from a **Lark** (international) or **Feishu** (China domestic) group chat. When a user drops a PDF or image into the designated intake group, RecordEngine automatically downloads it, runs AI extraction, and files it under the configured contact and folder — no manual upload required.

|                      | Detail                                                    |
| -------------------- | --------------------------------------------------------- |
| Supported file types | PDF, images (JPG, PNG, etc.)                              |
| Platforms            | Lark (open.larksuite.com) and Feishu (open.feishu.cn)     |
| Trigger              | File or image message sent to the intake group            |
| Processing           | Auto-profile detection → AI extraction → outbound webhook |
| Configuration        | Settings → General → 🐦 Lark / Feishu Integration         |

***

## Prerequisites

* A RecordEngine instance with a public HTTPS URL
* **Server URL** configured in Settings → General → 🔗 Server URL
* A Lark or Feishu developer account at [open.larksuite.com](https://open.larksuite.com) or [open.feishu.cn](https://open.feishu.cn)
* At least one Contact and Folder already created in RecordEngine

***

## How It Works

```
User drops file in Lark/Feishu group
        ↓
Lark/Feishu POSTs event to RecordEngine
        ↓
RecordEngine responds 200 OK immediately
        ↓  (background)
RecordEngine downloads file from Lark/Feishu API
        ↓
File saved + document record created in RecordEngine
        ↓
xr-watcher picks up → auto-profile detection → AI extraction
        ↓
Outbound webhook fires → Make.com → Google Sheets / notifications
```

RecordEngine responds to Lark/Feishu instantly and processes the file in the background. This prevents Lark/Feishu's 3-second timeout from triggering a retry and creating duplicate documents.

***

## Step 1 — Create a Lark / Feishu App

1. Go to your developer console:
   * **Lark:** [open.larksuite.com/app](https://open.larksuite.com/app)
   * **Feishu:** [open.feishu.cn/app](https://open.feishu.cn/app)
2. Click **Create App** → **Custom App**
3. Give it a name (e.g. `RecordEngine Intake`) and save
4. Go to **Credentials & Basic Info** and note your:
   * **App ID** — looks like `cli_xxxxxxxxxxxxxxxx`
   * **App Secret**

***

## Step 2 — Grant App Permissions

In your app's developer console, go to **Permissions & Scopes** and enable:

| Permission              | Purpose                             |
| ----------------------- | ----------------------------------- |
| `im:message`            | Read messages sent to the app       |
| `im:message.receive_v1` | Receive message events              |
| `im:resource`           | Download file and image attachments |

Save and publish a new app version after adding permissions.

***

## Step 3 — Configure the Webhook URL

1. In your developer console, go to **Events & Callbacks → Event Configuration**
2. Enable events and set the **Request URL** to:

```
https://yourinstance.recordengine.ai/api/lark/webhook
```

<Note>
  Your exact webhook URL is shown in RecordEngine under **Settings → General → 🐦 Lark / Feishu Integration → Webhook URL**. Copy it directly from there.
</Note>

3. Subscribe to the event: `im.message.receive_v1`
4. Save. Lark/Feishu will send a challenge request to verify your URL — RecordEngine handles this automatically.

***

## Step 4 — Configure RecordEngine Settings

In RecordEngine, go to **Settings → General** and scroll to **🐦 Lark / Feishu Integration**.

| Field              | Description                                           |
| ------------------ | ----------------------------------------------------- |
| **Platform**       | Select **Lark (International)** or **Feishu (China)** |
| **App ID**         | Paste your App ID from Step 1                         |
| **App Secret**     | Paste your App Secret from Step 1                     |
| **Intake Contact** | Documents will be filed under this contact            |
| **Intake Folder**  | Documents will be placed in this folder               |
| **Intake Profile** | AI extraction profile applied to incoming documents   |
| **Webhook URL**    | Read-only — copy this into your developer console     |

Click **💾 Save Lark Settings** when done.

<Tip>
  Choose a dedicated contact and folder for Lark intake (e.g. a contact named "Lark Intake" with a folder named "Inbox"). This keeps incoming documents organised and easy to review before filing them elsewhere.
</Tip>

***

## Step 5 — Add RecordEngine Bot to Your Intake Group

1. In Lark/Feishu, open the group you want to use as the intake channel
2. Go to group **Settings → Bots → Add Bot**
3. Search for your app name (e.g. `RecordEngine Intake`) and add it

From this point on, any **file or image** sent to this group will be picked up by RecordEngine. Text messages are silently ignored.

***

## Step 6 — Publish Your App

In the developer console, go to **App Release → Version Management** and publish your app. RecordEngine will not receive events until the app is in **Published** status.

***

## Testing the Integration

1. Send a PDF or image to your intake group in Lark/Feishu
2. Within a few seconds, the document should appear in RecordEngine under the configured contact and folder with status **Processing**
3. Once AI extraction completes, status changes to **Ready** (or **Needs Review** depending on confidence score)

If you have an outbound webhook configured, a notification will also be sent to your downstream system (e.g. Google Sheets, Lark Base).

***

## Lark vs Feishu

Lark and Feishu are the same platform — Lark is the international version, Feishu is the China domestic version. The only difference is the API base URL:

| Platform             | API Base URL                 |
| -------------------- | ---------------------------- |
| Lark (International) | `https://open.larksuite.com` |
| Feishu (China)       | `https://open.feishu.cn`     |

All API paths are identical. RecordEngine switches automatically based on your **Platform** setting.

***

## Supported Message Types

| Message type                  | Handled    |
| ----------------------------- | ---------- |
| File (PDF, Word, Excel, etc.) | ✅ Yes      |
| Image (JPG, PNG, etc.)        | ✅ Yes      |
| Text message                  | ⏭️ Ignored |
| Audio / Video                 | ⏭️ Ignored |

***

## Troubleshooting

**Document not appearing in RecordEngine after sending a file**

* Confirm your app is **Published** (not in Draft) in the developer console
* Confirm the **Request URL** in Events & Callbacks matches your RecordEngine webhook URL exactly, including `https://`
* Confirm **App ID** and **App Secret** are saved correctly in RecordEngine Settings
* Check that the RecordEngine bot has been added to the intake group

**Webhook URL verification failed in Lark/Feishu developer console**

* Confirm your RecordEngine instance is publicly accessible over HTTPS
* Confirm the **Server URL** is set in **Settings → General → 🔗 Server URL**
* The challenge verification is handled automatically — no additional configuration is needed

**Documents arriving twice**

* Check whether any Make.com scenario is also forwarding Lark events to RecordEngine. If so, deactivate it — RecordEngine handles the full intake flow internally without Make.com involvement.

**File downloaded but extraction profile is wrong**

* Go to **Settings → General → 🐦 Lark / Feishu Integration** and update the **Intake Profile**
* For mixed document types, consider using the **Universal Invoices and Receipts** profile, or enable auto-profile detection

***

## Related

* [API-Mode Deployment](/docs/api-mode-deployment)
* [Extraction Profiles](/docs/extraction-profiles)
* [Outbound Webhooks](/docs/webhooks)