SKILL.md
Editor — Captions
Add animated captions through the external ZapCap API. Keep ZapCap credentials, billing, and task state separate from the Vidjutsu tenant.
ZapCap API
- Base URL:
https://api.zapcap.ai - Auth header:
x-api-key: <ZAPCAPAPIKEY>
Step 1 — Upload video
POST https://api.zapcap.ai/videos/url
x-api-key: <ZAPCAP_API_KEY>
{ "url": "[VIDEO_URL]" }Returns { "id": "..." }.
Step 2 — Create captioning task
POST https://api.zapcap.ai/videos/{id}/task
x-api-key: <ZAPCAP_API_KEY>
{
"templateId": "[TEMPLATE_ID]",
"language": "en"
}Returns { "taskId": "..." }.
Step 3 — Poll until complete
GET https://api.zapcap.ai/videos/{id}/task/{taskId}
x-api-key: <ZAPCAP_API_KEY>Poll every 5 seconds. When status is "completed", the response includes a downloadUrl.
Step 4 — Download result
Fetch the downloadUrl to get the captioned video.
Parameters
| Field | Type | Default | Description |
|---|---|---|---|
templateId |
string | ZapCap default | Caption style template ID |
language |
string | auto-detect | Language code (e.g., "en", "es") |
Pipeline Integration
Captions should be applied after overlays (ZapCap renders on top of everything):
1. Generate video (any director-* skill)
2. vidjutsu overlay — burn text overlay (optional)
3. Upload overlaid video URL to ZapCap → poll → download captioned resultKey Behaviors
- Input must be a public URL — if you have a local file, upload it via
vidjutsu upload <file>first to get a URL. - Apply captions last — ZapCap renders on top of everything.
- Poll, don't block — captioning takes 30-120 seconds typically.
- ZapCap billing is separate — it is not part of the Vidjutsu subscription.
- Keep task authority separate — ZapCap's
taskIdis used only for this
external caption job and must never be confused with a Vidjutsu clone task ID.