How to Build an AI Face‑Swap Workflow with Coze in 7 Simple Steps
This step‑by‑step guide shows how to create a Coze workflow, add and configure start, face‑swap, and end nodes, test the workflow with sample image URLs, publish it, retrieve its workflow ID, and invoke the Coze API using a JSON payload.
1. Create workflow
In Coze, start a new workflow.
2. Add nodes
Insert the “Intelligent Face Swap” node from “More Image Plugins”.
3. Configure nodes
3.1 Start node
Define two input parameters: source_url (URL of the source portrait) and target_url (URL of the target portrait).
3.2 Face‑Swap node
Map the start node outputs to the face‑swap node inputs: reference_picture_url ←
source_url template_picture_url← target_url The node returns the swapped image.
3.3 End node
Select the node’s return value as the workflow output.
4. Test run
Use two publicly accessible portrait images, for example:
Source: https://qcloud.dpfile.com/pc/6it6-wSAtuDIQ4I0sC7Mwl-38nWSg4HL9q5EUval_2sJ7OUzNmXLyzmJmWen46VO.jpg
Target: https://img1.baidu.com/it/u=3021032973,104136814&fm=253&fmt=auto&app=138&f=JPEG?w=500&h=889
5. Publish workflow
6. Retrieve workflow ID
The generated ID is 7530470053024759808.
7. Call Coze API
Send a POST request to the Coze execution endpoint with the authentication token, workflow_id, and a JSON payload containing source_url and target_url.
{
"source_url": "https://qcloud.dpfile.com/pc/6it6-wSAtuDIQ4I0sC7Mwl-38nWSg4HL9q5EUval_2sJ7OUzNmXLyzmJmWen46VO.jpg",
"target_url": "https://img1.baidu.com/it/u=3021032973,104136814&fm=253&fmt=auto&app=138&f=JPEG?w=500&h=889"
}AI Software Product Manager
Daily updates of Xiaomi's latest AI internal materials
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.
