curl --request POST \
--url https://developer.creao.ai/v1/runs \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '
{
"agent_id": "00000000-0000-4000-8000-000000000000",
"input": {
"company": "Apple",
"year": "2025"
}
}
'const url = 'https://developer.creao.ai/v1/runs';
const options = {
method: 'POST',
headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
body: JSON.stringify({
agent_id: '00000000-0000-4000-8000-000000000000',
input: {company: 'Apple', year: '2025'}
})
};
fetch(url, options)
.then(res => res.json())
.then(json => console.log(json))
.catch(err => console.error(err));import requests
url = "https://developer.creao.ai/v1/runs"
payload = {
"agent_id": "00000000-0000-4000-8000-000000000000",
"input": {
"company": "Apple",
"year": "2025"
}
}
headers = {
"Authorization": "Bearer <token>",
"Content-Type": "application/json"
}
response = requests.post(url, json=payload, headers=headers)
print(response.text){
"run_id": "11111111-1111-4111-8111-111111111111",
"status": "pending",
"agent_id": "00000000-0000-4000-8000-000000000000",
"conversation_id": "22222222-2222-4222-8222-222222222222",
"model": "anthropic/claude-haiku-4-5"
}Create an async run
Start a background CREAO CLI run.
curl --request POST \
--url https://developer.creao.ai/v1/runs \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '
{
"agent_id": "00000000-0000-4000-8000-000000000000",
"input": {
"company": "Apple",
"year": "2025"
}
}
'const url = 'https://developer.creao.ai/v1/runs';
const options = {
method: 'POST',
headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
body: JSON.stringify({
agent_id: '00000000-0000-4000-8000-000000000000',
input: {company: 'Apple', year: '2025'}
})
};
fetch(url, options)
.then(res => res.json())
.then(json => console.log(json))
.catch(err => console.error(err));import requests
url = "https://developer.creao.ai/v1/runs"
payload = {
"agent_id": "00000000-0000-4000-8000-000000000000",
"input": {
"company": "Apple",
"year": "2025"
}
}
headers = {
"Authorization": "Bearer <token>",
"Content-Type": "application/json"
}
response = requests.post(url, json=payload, headers=headers)
print(response.text){
"run_id": "11111111-1111-4111-8111-111111111111",
"status": "pending",
"agent_id": "00000000-0000-4000-8000-000000000000",
"conversation_id": "22222222-2222-4222-8222-222222222222",
"model": "anthropic/claude-haiku-4-5"
}Authorizations
Account API key from Developer Console. Keys start with cr_sk_.
Body
Active personal CREAO agent to run.
JSON input for the agent. Serialized input must be 256 KB or smaller.
Model id to run with (for example, anthropic/claude-haiku-4-5). Omit to use the default model, currently anthropic/claude-sonnet-5-5. Supported ids: https://agent.docs.creao.ai/developer-api/supported-models. Unknown ids return 400 MODEL_NOT_SUPPORTED; models not available on your plan return 403 MODEL_NOT_ALLOWED.
100Existing conversation to continue. Omit to create a new conversation.
Personal workspace whose root uploads the agent can read. On a new conversation this value is stored. On an existing conversation it must match the bound workspace or the conversation must be unbound. Mismatch returns 409 WORKSPACE_MISMATCH. Not accepted on agent create or edit runs.
Public http or https URL for the terminal async result.
2048Optional boolean gate for completion notifications. true makes this conversation eligible; false mutes it. Omit to inherit the conversation or account default. This is not an internal bitmask. Not accepted on realtime runs.
Response
Run accepted.
CREAO CLI run ID.
pending, running, completed, failed, cancelled Agent that ran. Creation runs return null until the agent is created, then the completed run includes the new agent ID.
Conversation ID for follow-up runs.
Model id this run executed with. null only for runs created before model selection shipped.
Personal workspace bound to this conversation, when set.
Final agent output. Completed runs commonly include text; agent-creation runs also include agent_id.
Show child attributes
Show child attributes
Settlement credits plus an optional per-model breakdown for the run.
Omitted when the run incurred no cost and collected no model usage.
models is present only for Super Agent–allowlisted accounts.
Token fields appear only for token-metered capabilities. Per-image
and per-second capabilities use generated_images or duration_seconds
instead. kind tells the client which fields apply.
Show child attributes
Show child attributes
Show child attributes
Show child attributes