Base Connectivity
Learn how to generate the base connectivity for your connector. Base connectivity is the foundation of your connector and it defines the basic structure of your connector.
The base connectivity includes:
-
{projectname}-connectivity-modelThis is the canonical connectivity model (CCM). This package contains the code that directly reflects the structure of the API specification from which a project is generated. This is written in LinkWeave, and is the base from which all connectors will be built.
-
{projectname}-{platformname}-connector-modelThis package represents a platform-specific connector, such as for Anypoint Platform, which takes the connectivity model as a base and adds customizations and transformations to address platform limitations or enhance the user experience. It also contains the
Connector.dwlfile, which defines the connector.
You can generate base connectivity either from an API specification (OAS 3.0) or with AI assistance (available when using Cursor).
Generate Base Connectivity from an API Specification
-
Open the Connector Builder extension in VS Code or Cursor.
-
Select New Project from API Spec.
-
Enter the required information in Project Details.
Field Description Name
Enter a name for your project. The name must be lowercase, alphanumeric, and contain no spaces.
Description
Enter a description for your project.
-
Optionally, click Select Icon to upload an icon for your project.
-
In Connectivity Platforms, select the Anypoint Platform checkbox.
-
Click Next.
-
Click Select API Spec file and then select the API specification (OAS 3.0) to upload to create your connector.
-
Click Open Spec.
-
Select the API endpoints that you want your project to include.
-
Click Next.
-
Select the folder where your project is contained. A subfolder is automatically created within this folder.
-
Click Select containing folder. Your project is created, and VS Code opens the newly created project.
Generate Base Connectivity with AI Assistance (Cursor)
-
Open the Connector Builder extension in Cursor.
-
Select Settings and ensure Enable Ai Assisted Workflows is enabled.
-
Select New Project with AI.
-
Enter the required information in AI Project Details.
Field Description Name of SaaS Platform
Enter the name of the SaaS platform you want to create a connector for.
Top Use Cases
Enter the top use cases for the connector, from most important to least important. Use natural language to describe the use cases. Keep the use cases succinct, such as
List invoices,Create an invoice, orCreate an order. If you want full CRUD (Create, Read, Update, Delete) behavior, you must explicitly mention each one.Official API Documentation or Reference Sources
Enter the URLs of the official API documentation or reference sources for the connector.
Project Folder
Select the root directory of your project. All project files will be created here (a subfolder will not be created in this case).
-
Click Submit. This copies a prepared prompt to your clipboard, which you can manually paste into the Cursor chat window.
-
Open the Cursor chat window (if not already open) and paste the copied prompt.
-
Optionally (but recommended), copy your provided URLs from the
apiDocssection of the prompt and paste them just before the final line. Be sure to add an@symbol in front of each URL so that Cursor recognizes them as resources. When done correctly, the URLs are highlighted in blue. -
Ensure you’re in Agent mode. You can optionally choose which LLM that Cursor sends your prompt to.
Connector Builder has been tested with claude-sonnet-4. Each model performs differently and might yield different results. -
Click Send.
-
Wait for the agent to complete its work. AI by nature is non-deterministic, so running the same prompt multiple times might yield different results each time. The expected outcome is:
-
{projectroot}/discovery/discovery.jsonis generated, capturing the LLM’s recommendations for implementing the requested use cases. -
{projectroot}/discovery/{saasname}.yamlis created, containing the discovered endpoints, required schemas, and related details. -
The extension detects the new API specification and automatically generates a project based on it. This process may take several seconds.
-
Cursor closes and reopens in the root directory of the newly created project.
-



