Data Connectors

Data connectors are the unified entry point for managing external data sources. By creating connectors, applications can securely access enterprise databases, document systems, and object storage, querying and referencing data in real time during conversations.

Connector types

Connectors are divided into platform-managed and stream-processing types based on data storage and access patterns:

Category	Connector	Data storage	Use case
Platform-managed	File	Platform storage or self-owned OSS	Upload and manage unstructured documents (PDF, Word, Markdown, etc.)
Platform-managed	Spreadsheet	Platform storage or self-owned OSS	Import and query structured spreadsheet data (CSV, Excel, etc.)
Stream processing	MySQL	Data stays in source database, real-time access	Connect to MySQL databases, execute SQL queries
Stream processing	PostgreSQL	Data stays in source database, real-time access	Connect to PostgreSQL databases, execute SQL queries
Stream processing	PolarDB-X 2.0	Data stays in source database, real-time access	Connect to Alibaba Cloud PolarDB-X 2.0 distributed databases

Yuque and OSS are not standalone connector types. They serve as sync data sources within file connectors — see Sync Rules.

Prerequisites

Account permissions: Primary account or RAM user with data connector management permissions.
Data source preparation (by connector type):

Connector	Requirements
File / Spreadsheet	Files to upload, or an existing OSS Bucket
MySQL	MySQL database instance (Alibaba Cloud RDS or self-hosted), network accessible
PostgreSQL	PostgreSQL instance with `wal_level` set to `logical`
PolarDB-X 2.0	Alibaba Cloud PolarDB-X 2.0 instance in a region supporting private network access

Create a connector

Go to Data Ingestion → Data Connectors in the console, click Create Connector, select a type, and configure as follows.

File
Spreadsheet
MySQL
PostgreSQL
PolarDB-X 2.0

File connectors manage unstructured documents (PDF, Word, etc.).

Select File as the connector type.
Enter the Connector Name and Description. The description guides application accuracy — describe the data content and purpose.
For Storage Location, choose Platform Storage (up to 200,000 files / 1 TB, free for a limited time) or Self-owned OSS (first-time use requires authorization; add the bailian-connector-access tag with value ReadAndWrite to the target Bucket).

MySQL connectors are stream-processing type, connecting to MySQL databases.

Select MySQL as the connector type.
Enter the Connector Name and Description.
For Data Source, choose Create Custom Data Source (supports Alibaba Cloud RDS MySQL via SLR authorization or self-hosted MySQL with manual address/port) or Import from DMS (quickly import existing DMS data sources; only DMS-imported sources support SQL query execution).
Select Network Type: Public (add the specified IP range to the database whitelist) or Private (select region; more stable and secure for production).
Fill in the database connection details:

Field	Description
Database instance	Only for RDS MySQL; enter RDS instance ID (e.g., `rm-******adq7`)
Database address	Auto-filled for RDS MySQL; manual for self-hosted
Database port	Auto-filled as 3306 for RDS MySQL; manual for self-hosted
Database username	Required, for authentication
Database password	Required, ensure the user has read permissions

(Optional) Click Test Connection to verify connectivity. After successful detection, select the target database from the dropdown.

PostgreSQL connectors support Alibaba Cloud RDS PostgreSQL or self-hosted PostgreSQL.Additional prerequisites:

Database account must have high privileges (Superuser or REPLICATION permission)
Set wal_level to logical (default is replica)
For self-hosted: configure listen_addresses to allow 100.64.0.0/16 access

Select PostgreSQL as the connector type.
Enter the Connector Name and Description.
For Data Source, choose Create Custom Data Source (same as MySQL).
Fill in the database connection details:

Field	Description
Host	Database instance address
Port	Default 5432
Database name (dbName)	Required, specify the database to connect to
Username	High-privilege database account
Password	Database password

Click Test Connection. PostgreSQL uses DTS (Data Transmission Service) for connectivity detection.

MySQL vs PostgreSQL differences:

Difference	MySQL	PostgreSQL
Default port	3306	5432
Extra required field	None	dbName required
Connectivity detection	EventBridge	DTS
Special config	None	`wal_level` set to `logical`

Click Confirm to complete creation.

Import data

After creating a connector, import data into it.

File connector

Click Details on the target connector in the data connector list to open the Files tab.
Select an existing category from the left tree, or click + to create a new one.
Click Import Data, select Local Upload to upload files, or select OSS to pull files from OSS.
Select a parsing method (default or custom); the system handles parsing automatically.

JSON, CSV, and YAML formats are not directly supported. Convert to XLSX or XLS before importing.
Imported files are stored as independent copies, with no link to the original data.

Database connector

Database connectors (MySQL / PostgreSQL / PolarDB-X 2.0) keep data in the source database — no manual import needed. Once created, applications can query data in real time.

Sync rules

File-type connectors support sync rules for automatically syncing files from external sources. In the connector detail page's Files tab, click Sync Data Rules → Create Sync Rule.

OSS
Feishu
DingTalk
SharePoint
Yuque

Syncs OSS bucket data to the platform via Alibaba Cloud Service Linked Role (SLR).

OSS sync uses SLR to access your Bucket — no AccessKey needed. First-time use requires authorization.
Object paths must include the Bucket name, e.g., my-bucket/docs/ or my-bucket/docs/foo.md.
Target Bucket must have the bailian-datahub-access tag (value read).

Parameter	Required	Description
OSS Region	Yes	Region of the OSS Bucket
OSS object path	Yes	Path including Bucket name (e.g., `my-bucket/docs/` for a directory, or `my-bucket/docs/foo.md` for a single file)
Sync period	Yes	Auto-sync frequency
Connection test	Yes	Verify the OSS path is accessible
Data tags	No	Classification tags, max 32 chars each

Syncs Feishu knowledge bases and documents via app credentials.

Create a self-built app on the Feishu Open Platform to obtain App ID and App Secret.
Grant wiki, cloud document, spreadsheet read permissions and contact user ID read permission to the app.
One of the following must be provided: knowledge base ID, directory ID, or document ID. Supports docs, spreadsheets, bitable, and PPT nodes. Local uploads are not supported.
Feishu export permissions must be enabled. Mindnote documents are not supported.

Parameter	Required	Description
Feishu App ID	Yes	App ID of the self-built app
Feishu App Secret	Yes	App Secret
Feishu knowledge base ID	One of three	Knowledge base ID, directory ID, or document ID — one required
Feishu directory ID	One of three	Same as above
Feishu document ID	One of three	Same as above
Excluded document IDs	No	List of Feishu document IDs to exclude from sync
Feishu document domain	No	Domain prefix (e.g., `xxx` from `xxx.feishu.cn/wiki/...`)
Sync period	Yes	One minute / One hour (default) / One day
Connection test	Yes	Verify app credentials and knowledge base access
Data tags	No	Classification tags, max 32 chars each

Syncs DingTalk knowledge base documents and spreadsheets via app credentials and MCP services.

Create an app on the DingTalk Open Platform to obtain Client ID and Client Secret.
Required permissions: Contact.User.Read, Wiki.Workspace.Read, Wiki.Node.Read, Document.Workbook.Read, Notable.Base.Read_All.
One of the following must be provided: knowledge base ID, folder ID, or document ID.
Enable DingTalk Document MCP and DingTalk Spreadsheet MCP on the DingTalk MCP Marketplace, and obtain the MCP Endpoint links.
Supports syncing DingTalk documents, spreadsheets, and AI tables. Local uploads are not supported.

Parameter	Required	Description
DingTalk user ID	Yes	Obtain from Contacts → Member Management
DingTalk App ID	Yes	Client ID
DingTalk App Secret	Yes	Client Secret
DingTalk knowledge base ID	One of three	Knowledge base ID, folder ID, or document ID — one required
DingTalk folder ID	One of three	Same as above
DingTalk document ID	One of three	Same as above
DingTalk document MCP link	Conditional	Required when syncing DingTalk documents
DingTalk spreadsheet MCP link	Conditional	Required when syncing DingTalk spreadsheets
Sync period	Yes	Auto-sync frequency

Each sync consumes DingTalk Open Platform API quota, split into pre-sync validation and sync execution phases. Evaluate quota capacity in advance.

Syncs SharePoint documents and folders via Azure AD app registration.

SharePoint sync is currently available to whitelisted users only. Contact the Alibaba Cloud Bailian team to request access.

Prerequisite: Register an Azure AD app

Sign in to the Azure portal, search for Microsoft Entra ID.
Under Management → App registrations, create a new app. Copy the Directory (tenant) ID and Application (client) ID.
Under Certificates & secrets, create a client secret and copy the value (shown only once).
Under API permissions, add Sites.Read.All and Files.Read.All (application permissions), and grant admin consent.

Parameter	Required	Description
Tenant ID	Yes	Azure AD app Directory (tenant) ID
Client ID	Yes	Azure AD app Application (client) ID
Application secret	Yes	Client secret value (shown only once at creation)
SharePoint share links	Yes	Share links, one per line
Cloud type	No	Global public cloud (default) or China (21Vianet)
Sync period	Yes	One minute / One hour (default) / One day
Connection test	Yes	Verify credentials and share links
Data tags	No	Classification tags, max 32 chars each

Syncs Yuque documents and knowledge bases via token-based access.Obtain a Yuque Token:

Personal Token: Sign in to Yuque Settings, create and copy a token.
Team Token: Navigate to Team Space → Team Settings → Developer → Tokens, create and copy a token.

Parameter	Required	Description
Yuque doc/knowledge base URL	Yes	Full URL list, comma-separated (`,`)
Yuque Token	Yes	Personal or team token
Document permission level	No	`0`: private, `1`: public, `2`: organization-visible. Multi-select supported (e.g., `0,2`), default `0,1,2`
Sync period	Yes	One minute / One hour (default) / One day
Connection test	Yes	Verify token and URL accessibility
Data tags	No	Classification tags, max 32 chars each

After configuration, click Connection Test to verify connectivity. Once passed, configure data tags and click Confirm to create the rule. The rule appears in the sync rules list, where it can be enabled or disabled.

File connector tools

File connectors automatically generate the following tools, callable from Agent applications or workflows:

Tool	Description
`searchFile`	Search files by keyword, returns matching file list with download links
`getFile`	Retrieve file by ID, returns a download link

Manage connectors

From the data connector list, you can perform the following operations on existing connectors:

Operation	Description
Edit	Modify connector name and description only
Duplicate	Copy connector configuration to create a new connector
Delete	Delete the connector — irreversible

Once a connector is bound to a knowledge base as a data source, the binding cannot be changed. To switch data sources, create a new knowledge base.

Use connectors in Agents

In the Agent editor, go to Skills → Data Connectors.
Click + and select the desired data connector.
Click Publish — the Agent can now invoke the connector's tools in conversations.

Use connectors in workflows

In the workflow editor, add a Data Connector node.
Select a connector and the tool to use (one tool per node).
The tool outputs a result object containing content (Array) and isError (Boolean) fields.

Data Import API

Beyond the console, you can manage ingestion through the API — see API Reference — Data Import.

Data Import endpoints (/api/v1/connector/dash/) share the same workspace URL and authentication as all other endpoints — see API Overview.

With a connector configured, move on to Create a Knowledge Base to build a searchable index from the connector's data.

​Connector types

​Prerequisites

​Create a connector

​Import data

​File connector

​Database connector

​Sync rules

​File connector tools

​Manage connectors

​Use connectors in Agents

​Use connectors in workflows

​Data Import API

Connector types

Prerequisites

Create a connector

Import data

File connector

Database connector

Sync rules

File connector tools

Manage connectors

Use connectors in Agents

Use connectors in workflows

Data Import API