API Reference · Agent Tasks

Wake the workload's AI daemon from idle sleep

Idle AI daemons scale to zero after ~30 minutes without chats or tasks. This scales the daemon back up (a no-op when it is already running) and waits briefly for its agent API to answer. Mostly for automation that talks to a daemon out-of-band — e.g. an AI master's fleet tools call it before delegating to a slave agent; the platform's own chat and task endpoints wake the daemon implicitly. No body.

POST/v1/tenants/{name}/workloads/{id}/agent-wake

Authorization

x-api-keystringrequired
Your per-workspace API key, e.g. llc_AbC123…. Also accepted as Authorization: Bearer <key>. Mint keys on your workspace's API keys page — each key belongs to one workspace.

Path parameters

namestringrequired
Workspace (tenant) name.
idstringrequired
Workload id within the workspace.

Response

200Wake accepted.

statusenum
`awake` — the daemon answered and is ready now. `waking` — the pod is still starting; poll or retry in a few seconds.
curl -X POST https://api.live-llm.com/v1/tenants/acme/workloads/web/agent-wake \
  -H "x-api-key: $LIVELLM_KEY"
Try itPOSThttps://api.live-llm.com/v1/tenants/{name}/workloads/{id}/agent-wake
Fill required path params: name, id.
Example response200
{
  "status": "awake"
}