Building your connector using AI

This guide explains how to build a Keragon connector using our AI agent. You provide your API documentation, a short list of actions and triggers, and optionally some test credentials, and our AI agent generates the connector. The AI agent is the right path when you need a connector live fast and don't want to tie up engineering.

 

 

Getting started

Log in to the partner platform and navigate to the Connectors page. 
If you don't have any connectors yet, click Set up your connector. If you already have existing connectors, click Create new.

Step 1: Name your connector

Enter your official product name. This is the name Keragon users will see in the marketplace, so use the real name your customers know you by.

Step 2: Choose how to build

After entering your product name, you'll see two options:

  • Let our AI agent build it: Share your API docs and let AI do the heavy lifting. 
  • Build it yourself: Your team builds the connector using our CLI.

Select Let our AI agent build it to continue with the AI-assisted flow.

Step 3: Provide your details

The AI builder asks for three pieces of information.

  1. API documentation (Required)
    This is the most important piece of information we need. Paste a link to your API reference docs so the AI can understand your endpoints, request/response formats, and error handling.
    If you don't have a public URL, click "Don't have a URL? Paste the docs instead" and provide the text directly.
    Without API documentation, the agent cannot build the connector.
     
  2. Actions and triggers (Required)
    List the actions and/or triggers you want the connector to support:
    • Actions are things your connector does — for example, "Create patient" or "Send message."
    • Triggers are events coming from your product that Keragon listens for — for example, "New appointment created" or "Document signed." When something happens on your side, Keragon picks it up.
      Keep it small; two to five is the sweet spot. The AI does a better job when it has less to build at once. You don't need to map every endpoint in your API upfront. You can always come back and add more using the same AI flow later, and Keragon users can request new actions and triggers from their dashboard at any time and we'll add them at no extra cost.
      Not sure how to describe your actions and triggers? Click "Not sure what to write? See guide" for examples and best practices.
       
  3. Test credentials
    Partners who provide test credentials typically go live 3x faster.
    These should be credentials for a sandbox or test account; not production credentials, and not anything that provides access to real patient data. Just a test environment we can use to verify the connector works end to end.
    Test credentials are stored securely on our side. Select your authentication type from the dropdown and provide the necessary details.
     

Step 4: Click "Build"

Once you've filled in your details, click Build with AI and the agent will start generating your connector.
 



What happens after you click build

The connector doesn't go live immediately. After the AI builds the connector, our team will:

  • Review the generated code
  • Test the connector against your API 
  • Make any adjustments needed
  • Reach out if we need to clarify anything about your API or use cases

By the time your connector goes live, it has been verified by both the AI and our team.
 

Editing existing connectors

You can use the same AI flow to update connectors that are already live. From the Connectors page, open your connector and use the AI builder to add new actions or triggers, update your authentication scheme, or make other changes. The same review process applies.
 

Was this article helpful?
0 out of 0 found this helpful

Articles in this section