Last week we conducted research on suitable alternatives to Voiceflow. This was because Voiceflow has no API access for developers - it only allows for exporting conversations into a visual representation (PDF/PNG) which doesn't help us extract the conversation data at all.


Having looked at some alternative platforms, we got in contact with the client to hear their opinion on our research. The client indicated that of all the platforms we suggested, Voiceflow is still the best and we should investigate further to see if we are able to get data out of it.

What are we looking to achieve?

The client has asked us to ignore alternatives and continue working with Voiceflow for now.

1. Voiceflow Diagram Extraction We need to find a solution for extracting the conversation data from the Voiceflow.

Conversation JSON

Later in the week, we conducted a closer analysis of Voiceflow and discovered that a JSON file is generated when the chatbot is uploaded to Amazon Alexa or Google Assistant. The JSON file has all the information we need about the conversation, including output speech and possible branches. It stores all the nodes (speech blocks) in the conversation as well as the links between them.

		**"nodes"** : {
				"5f9f562f11b31520c2a2a6eb" : {
						"nodeID" : "5f9f562f11b31520c2a2a6eb"
		**"links"** : {
				"3f8f437e12c05412c6d5d4eb" : {
						"source" : "5f9f562f11b31520c2a2a6eb",
						"destination" : "5fa30d3eae56a9feaddc7e3c"

New Goals:

It should be mentioned, however, that there are a few issues with the diagrams generated.

Voiceflow Diagram Problems