Motion typography & posters — POST /v1/typography

For agents that compose their own video (HTML, After Effects, any renderer). WideCast only designs the text, with the same engines the editor uses: template typography for a key phrase mid-video, and the SVG poster engine for the opening hook poster and the closing call-to-action poster. You get every object back as its own self-contained SVG — the text already converted to outlines with the fonts WideCast renders with, so it needs no font and stays sharp at any size — plus a transparent PNG, its text and its position. Charges 1 credit per call.

Line breaks. In quote, | (also ->, → or a new line) forces a line break: "KHÔNG CẦN QUAY | VẪN CÓ KHÁCH" sets two lines.

Timing, motion and sound are yours. The response carries the design only. Reveal each object on the word that says it, animate it in your video's own style, and add sound effects to match your motion.

Safe area. Every object stays inside y 128–960 of the 1280 canvas: clear of the platforms' dead zones (the top 10 % and the bottom 25 % of a vertical video).

Request

curl -sS -X POST "https://widecast.ai/app/dashboard/v1/typography" \
  -H "Authorization: Bearer wc_live_REPLACE_ME" \
  -H "Content-Type: application/json" \
  -d '{
        "role":  "typography",
        "text":  "Không cần quay video mà vẫn có khách hàng mỗi ngày",
        "quote": "KHÔNG CẦN QUAY | VẪN CÓ KHÁCH",
        "language": "vi",
        "look":  "my-video-01",
        "wait":  60
      }'

Field reference

Response

{
  "object": "typography",
  "id": "typo_5805cf4ae9c74941",
  "status": "completed",
  "role": "typography",
  "look": "my-video-01",
  "lines": ["KHÔNG CẦN QUAY", "VẪN CÓ KHÁCH"],
  "canvas": { "width": 720, "height": 1280, "safe_top": 128, "safe_bottom": 960 },
  "preview_url": "https://widecast.ai/downloads/<company>/api_typography/my-video-01/typo_5805cf4ae9c74941_preview.png",
  "objects": [
    { "index": 1, "name": "p0_deco1", "kind": "decoration", "text": null,
      "svg_url": "…_obj01.svg", "image_url": "…_obj01.png",
      "box": { "x": 60, "y": 430, "w": 600, "h": 12 }, "phase": 0 },
    { "index": 2, "name": "p0_line4", "kind": "text", "text": "KHÔNG CẦN QUAY",
      "svg_url": "…_obj02.svg", "image_url": "…_obj02.png",
      "box": { "x": 67, "y": 462, "w": 586, "h": 81 }, "phase": 0 },
    { "index": 3, "name": "p0_line5", "kind": "text", "text": "VẪN CÓ KHÁCH",
      "svg_url": "…_obj03.svg", "image_url": "…_obj03.png",
      "box": { "x": 86, "y": 519, "w": 547, "h": 80 }, "phase": 0 }
  ],
  "credits_charged": 1
}

• objects are in paint order (back to front). Word-by-word layouts return one object per word, each with its own text.
• svg_url: the object alone, text in outlines, sized to its box. Use it in a browser-based renderer (an <img> or inline SVG); it scales without blur.
• image_url: the same object as a transparent PNG at twice the canvas scale (box.w×2 by box.h×2 pixels), for renderers that do not take SVG.
• box is where WideCast places the object on the 720×1280 canvas: multiply by 1.5 for 1080×1920. Keep the objects' relative positions to keep the design, or rearrange them for your frame.
• phase appears when the design has several phases (p0_, p1_ …) that replace each other over time.
• preview_url shows the whole set as WideCast places it: look at it before placing the objects.
• status is processing while the job runs: poll GET /v1/typography/{id} (free; pass ?wait= up to 100 s). On failure status is failed and error says why.

Poll

curl -sS "https://widecast.ai/app/dashboard/v1/typography/typo_5805cf4ae9c74941?wait=30" \
  -H "Authorization: Bearer wc_live_REPLACE_ME"

MCP

widecast_generate_typography (waits up to ~90 s) and widecast_get_typography, on both the remote connector and the local widecast-mcp server.