Salesforce Integration Guide

Filter and route customer service tickets using IndieML via Salesforce Flow's native HTTP Callouts.

Prerequisite: Request API Key

Before starting, you need an API key to authenticate the connection. Click here to get your free API key. Copy the key and keep it handy for the configuration step.

Option 1: The Quick Install (Unmanaged Package)

You can deploy the complete Flow, External Service, and JSON schemas directly into your Salesforce environment using our unmanaged package.

1. Install the Package

  • Click here for the package link.
  • Log into your Salesforce environment.
  • Select Install for Admins Only and click Install.

2. Update the API Key

  • Go to the Setup page.
  • Navigate to Quick Find and search for Named Credentials.
  • Click on the IndieML credential.
  • Scroll down to the Principals section, click the arrow below Actions, and select Edit.
  • Parameter 1 > Value: Paste your API key here.
  • Click Save.

Your integration is now active. You can go to Setup > Flows to view and customize the IndieML Flow Blueprint.

Option 2: Build from Scratch

If you prefer to build the architecture yourself or want to understand exactly how the integration routes data, follow this step-by-step blueprint. You can set up the entire pipeline using only Flow Builder (no Apex code required).

Phase 1: The Secure Vault (Named Credentials)

1. Create the External Credential

  • Go to the Setup page
  • Navigate to Quick Find and search for Named Credentials.
  • Click the External Credentials tab and click New.
  • Label / Name: IndieML
  • Authentication Protocol: Custom
  • Click Save.

2. Set the Principal

  • From the IndieML page, scroll to Principals and click New.
  • Parameter Name: ApiKey
  • Sequence Number: 1
  • Under Authentication Parameters, click Add.
    • Name: ApiKey
    • Value: Paste your API key here.
  • Click Save.

3. Set the Custom Header

  • From the IndieML page, scroll to Custom Headers and click New.
  • Name: X-API-Key
  • Value: {!$Credential.IndieML.ApiKey}
  • Sequence Number: 1
  • Click Save.

4. Create the Named Credential

  • Go back to the Named Credentials page, click the Named Credentials tab, and click New.
  • Label: IndieML Endpoint
  • Name: IndieMLEndpoint
  • URL: https://indieml.app/v1/score/substance
  • Enabled for Callouts: Checked (ON)
  • Authentication:
    • External Credential: IndieML
  • Callout Options:
    • Generate Authorization Header: Unchecked (OFF)
    • Allow Formulas in HTTP Header: Checked (ON)
    • Allow Formulas in HTTP Body: Unchecked (OFF)
  • Click Save.
Phase 2: Granting User Access (Permissions)

Note: The user triggering the Flow must have access to the External Credential Principal, or the callout will fail silently.

  • Navigate to Quick Find and search for Profiles (or Permission Sets).
  • Select System Administrator (and any other profile that will trigger this Flow).
  • Click Enabled External Credential Principal Access and click Edit.
  • Select IndieML-ApiKey from the Available list and click Add to transfer it to the Enabled list.
  • Click Save.

Phase 3: The "Hello World" Sandbox Flow

1. Create the Callout Action

  • Navigate to Quick Find, search for Flows, and click New Flow.
  • Select Autolaunched Automations.
  • Select Autolaunched Flow (No Trigger).
  • On the blank canvas, click the + icon and select Action.
  • At the bottom of the right panel, click Create HTTP Callout.
  • Name the Callout: GetSubstanceScore
  • Named Credential: Select IndieML Endpoint from the dropdown, then click Next.
  • Configure Invocable Action:
    • Label: Score Text
    • Method: POST
    • Click Next.

2. Define the JSON Schemas

  • Under Sample JSON Request, paste:
Request Body 
{
  "input_text": "This is a test run."
}
  • Click Review, then click Next.
  • Under Select Sample Response Method, select Use Example Response and click Next.
  • Under Sample JSON Response, paste:
Response Body 
{
  "result": {
    "text": "This is a test run.",
    "substance_score": 0.18
  }
}
  • Click Review, then click Save.

3. Configure the Flow Variables

You will now be back on the Main Flow Builder screen.

  • Label: Call IndieML API
  • API Name: (Let this auto-populate to Call_Indie_ML_API)
  • Under Set Request Body > Value, click and select New Resource.
  • API Name: requestPayload (Data Type and Apex Class will auto-populate). Click Done.
  • Click the + icon on the canvas below Autolaunched Flow.
  • Select Logic > Assignment.
  • Label: Set Text To Score
  • API Name: (Let this auto-populate to Set_Text_To_Score)
  • Variable: Select your new requestPayload resource.
    • The menu will reset. Click requestPayload a second time to display the input_text option.
    • Select input_text from the newly expanded list.
    • Verify: The box should now show requestPayload > input_text.
  • Operator: Equals
  • Value: Click New Resource at the very bottom of the dropdown list.
    • Resource Type: Constant
    • API Name: textToScore
    • Data Type: Text
    • Value: The quick red fox jumps over the lazy dog.
  • Click Done.

Phase 4: Run the Debugger

  • Click Save in the top right corner.
  • Flow Label: IndieML Substance Flow
  • Flow API Name: (Let this auto-populate to IndieML_Substance_Flow)
  • Click Save.
  • Click Debug, then click Run at the bottom of the left side popout.
  • Look at the Debug Details in the left side popout.
  • Click Details under GetSubstanceScore.Score Text (External Services): Call IndieML API
  • Under 2XX, you should see a substance score of 0.19.
Addendum 1: Adapting the flow to a Live Pipeline

When moving from the autolaunched sandbox to a live production environment, keep these enterprise architecture rules in mind:

  • Use an Asynchronous Path: A Record-Triggered Flow cannot make an HTTP Callout on the immediate execution path. You must configure the Flow to run asynchronously (e.g., using a "Run Asynchronously" path) so the callout doesn't block the record save transaction.
  • Map Live Data: In your Assignment block, map the requestPayload > input_text variable directly to {!$Record.Description} to ingest live customer tickets.
  • Build a Fault Path: External APIs can timeout or return 401/500 errors. Always drag a Fault Path off your HTTP Callout element to handle exceptions gracefully (e.g., log an error record or send an alert) so the Flow doesn't crash the interview.
  • Evaluate the Score: Add a Decision block after the HTTP Callout to evaluate the score (e.g., Outputs from Call_Indie_ML_API > 2XX > result > substance_score is Less Than 0.30).
  • Manage Bulk Limits: Record-triggered flows execute an interview per record. If you mass-update hundreds of Cases via Data Loader, it will fire hundreds of individual HTTP callouts. Ensure your API rate limits (and Salesforce's outbound callout limits) can handle your expected bulk transaction volume.
Addendum 2: Modifying the Package for any External API

Because this unmanaged package deploys native, standard Salesforce metadata, you can use it as a foundational boilerplate to integrate any third-party REST API into Flow without writing Apex.

1. Install the Package

  • Click here for the package link.
  • Log into your Salesforce environment.
  • Select Install for Admins Only and click Install.

2. Update the API URL

  • Go to the Setup page.
  • Navigate to Quick Find and search for Named Credentials.
  • Click the Named Credentials tab.
  • Click the dropdown arrow next to IndieML Endpoint and select Edit.
  • Enter the new API base URL and click Save.

3. Update the API Key

  • Stay in Named Credentials, but click the External Credentials tab.
  • Click on the IndieML credential.
  • Scroll down to the Principals section, click the arrow below Actions, and select Edit.
  • Under Parameter 1, paste your new API key into the Value field.
  • Click Save.

4. Update the Custom Header (If Required)

Different APIs expect different authorization headers (e.g., Authorization vs X-API-Key).

  • From the IndieML External Credential page, scroll down to the Custom Headers section.
  • Click the dropdown arrow next to the existing X-API-Key header and select Edit (or delete it and click New).
  • Name: Enter the header name your new API expects (e.g., Authorization).
  • Value: Update the formula to match the expected format. For example, if your API requires a Bearer token, you would enter: Bearer {!$Credential.IndieML.ApiKey}.
  • Click Save.

5. Update the Request & Response Schemas

  • Navigate to Quick Find and search for External Services.
  • Under External Service Name find the generated service, click the arrow below Actions, and select Edit.
  • Replace the sample JSON request and response payloads with the schema expected by your new API. Salesforce will automatically regenerate the input and output variables.
  • Click Save.

6. Remap the Flow Action

  • Open the routing flow in Flow Builder.
  • Double-click the HTTP Callout action element on the canvas.
  • Map your Salesforce record values to the newly generated request fields.
  • Update any downstream Decision elements to reference the new output variables returned by your API.