### Answer To import a conversation flow agent via JSON, you can use the Dashboard's import agent feature. [(1) ]() When creating a conversation flow programmatically via the API, the JSON structure requires these key fields: Required fields: -

model_choice

- The model configuration (e.g.,

{ "model": "gpt-4.1", "type": "cascading" }

) [(2) ]() -

start_speaker

- Either

"user"

or

"agent"

-

nodes

- Array of node objects defining the conversation flow Node structure example:

json
{
  "id": "start",
  "type": "conversation",
  "instruction": {
    "type": "prompt",
    "text": "Greet the customer and ask how you can help them."
  },
  "edges": [
    {
      "id": "edge_1",
      "transition_condition": {
        "type": "prompt",
        "prompt": "Customer wants to book appointment"
      },
      "destination_node_id": "book_appointment"
    }
  ]
}

Optional fields include: -

global_prompt

- Prompt used across all nodes -

tools

- Array of available tools/functions -

start_node_id

- ID of the starting node -

default_dynamic_variables

- Variables referenced throughout the flow -

knowledge_base_ids

- For RAG integration -

components

- Local sub-flows embedded in the conversation Note: If importing via the Dashboard and using a custom voice, you may encounter a "failed to get voice id" error. [(1) ]()

​ To learn more, see the sources I considered: [[1] Conversation Flow: Import / Upload JSON](https://discord.com/channels/1186424376300679209/1394595851942367252) [[2] Create Conversation Flow](https://docs.retellai.com/api-references/create-conversation-flow) ​

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.