Pass PDF blob via HTTP Post Callout

I have an AWS lambda function, which helps me parse a PDF document to extract relevant data and fill a Salesforce object with this data afterwards in a Salesforce flow. I have a running setup, performing a HTTP POST callout using a pure (manually created) apex class and I would like to switch to the new Flow HTTP Callout element.

TL;DR

How can I pass a ContentVersion.VersionData base64 encoded blob of a pdf file to a Flow HTTP POST callout?

Apex approach

Using a ContentDocumentLink object (describing a PDF file / attachment linked to another object (e.g. a Case)) with a ContentDocumentId variable I retrieved a ContentVersion object by matching the ContentDocumentId on this object.

I pass this ContentVersionId of the ContentVersion object to an apex class, which retrieves the base64 encoded Blob of the PDF object like so

public static Blob get_content_version_blob_data(String ContentVersionId ) {  Blob pdfblob = [SELECT VersionData FROM ContentVersion WHERE Id = :ContentVersionId].VersionData;  return pdfblob;}

I then create the HttpRequest, performing the callout and get the response.

HttpRequest request = new HttpRequest();request.setEndpoint(<my_endpoint_url>);request.setMethod('POST');request.setHeader('Content-Type', 'pdf');request.setHeader('Accept-Charset', 'UTF-8');request.setBodyAsBlob(pdfblob);Http http = new Http();HttpResponse response = http.send(request);Map<String, Object> d_response = (Map<String, Object>) JSON.deserializeUntyped(response.getBody());

This approach is working perfectly fine, but with manual apex coding.

Hence I tried to transfer it to the new Flow HTTP Callout method.

Flow HTTP Callout approach

I created a simple flow with the following steps

1. Get ContentVersion
   • Retrieves the ContentVersion object, based on a match of the ContentDocumentId
   • The output is stored in a ContentVersion object named CurrentContentVersion, storing the variables ID (ContentVersionId) and VersionData (assumed to hold the blob data, as seen in the apex code)
2. Set callout body variable
   • The callout_body.body variable is assigned as the CurrentContentVersion.VersionData
3. Execute the HTTP Callout
   • As described in salesforce help Configure an HTTP Callout Action I created a new variable callout_body, using the Apex-Defined data type and referencing the automatically created apex class ExternalService__MyService_CalloutMyService_IN_body for my callout input body.
   • Note that in the JSON schema the request body of my ExternalService is defined as string
4. Assign the response to my local variables

JSON schema for request body in ExternalService

[...],"requestBody": {  "content": {    "application/json": {      "schema": {        "type": "object",        "properties": {          "body": {            "type": "string"          }        }      }     }  },  "required": true},[...]

Here are some screenshots of my flow

This approach already fails at step 2, assigning the callout_body.body variable with error message

Error Occurred: Value of class core.filemanager.FileBlobValue is not a ValueType: STRING

I can relate to this error message, since my JSON schema describes the body as type string. Hence I tried to change the JSON schema to type object for the body like

[...]"body": {  "type": "object"}[...]

But doing this, I cannot select the apex class ExternalService__MyService_CalloutMyService_IN_body any more for my callout_body input variable (see step 3), only the *_OUT_2XX and *_OUT_2XX_data options are available.

Same behaviour / issue when trying a binary input format for the body

[...]"body": {  "type": "string",  "format": "binary"}[...] 

Final question

Has anyone managed to define a blob style base64 encoded object type for the input body of a Flow HTTP POST callout to pass the blob to the external service?

Any help is very appreciated, thanks!