curl --request POST \
--url https://runchat.com/api/v1/{runchat_id} \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '
{
"runchat_instance_id": "existing-instance-uuid",
"inputs": {
"prompt_xyz123": "Generate a story about a robot.",
"maxTokens_abc456": [
100
]
}
}
'{
"metadata": "<string>",
"runchat_instance_id": "<string>",
"data": [
{
"id": "<string>",
"label": "<string>",
"type": "string",
"description": "<string>",
"data": [
{}
]
}
]
}{
"error": "<string>"
}{
"error": "<string>"
}Endpoints
Run
Trigger a Runchat workflow by ID. Pass inputs by parameter ID, or post a JSON body that flows into the workflow’s webhook node.
POST
/
{runchat_id}
curl --request POST \
--url https://runchat.com/api/v1/{runchat_id} \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '
{
"runchat_instance_id": "existing-instance-uuid",
"inputs": {
"prompt_xyz123": "Generate a story about a robot.",
"maxTokens_abc456": [
100
]
}
}
'{
"metadata": "<string>",
"runchat_instance_id": "<string>",
"data": [
{
"id": "<string>",
"label": "<string>",
"type": "string",
"description": "<string>",
"data": [
{}
]
}
]
}{
"error": "<string>"
}{
"error": "<string>"
}This endpoint runs a workflow that already exists. To build or edit a
workflow programmatically, the MCP server or
CLI are usually the easiest path — they expose
the full canvas toolset (create nodes, connect, run, publish) as native tools
and handle auth. The same tools are also available over plain HTTP via the
Canvas API.
Images and other media
- Inputs can be a URL or a base64 data URL
(
data:image/png;base64,...). Base64 inputs are uploaded to your gallery and replaced with a hosted URL before the workflow runs. - Outputs are returned as permanent
https://images.runchat.app/...URLs. Media generated on a provider’s CDN is copied to Runchat storage before the response is sent, so returned links don’t expire. Files stay until you delete them from your gallery.
Timing and credits
The request waits until the workflow finishes and returns all outputs in one response, so set a generous client timeout (around 10 minutes for slow models). Themetadata field is a JSON string; its credits object lists the credits
spent by usage type, summed across the nodes that ran.Authorizations
Body
application/json
Input data for the Runchat execution. Provide runchat_instance_id to continue an existing execution. The inputs object structure should match the parameters defined in the Runchat API schema, keyed by paramId_nodeId. Alternatively, send raw data for webhook-triggered flows.
Request body for executing a Runchat flow.
Response
Successful execution of the Runchat flow.