The exact script used in all four tests

To isolate the control, all four API requests used the same prompt, script, duration, voice, language, orientation, and fast scene pacing. Only the engine/style and onscreen_text_from_script_only value changed.

Small ideas become clear when we remove distraction. Start with one question, connect each step, and let every visual support the message. When the screen shows only words from this script, viewers can focus on the explanation.

Prompt: Create a concise visual explainer about turning complex ideas into clear stories.
Timing: 0.25 minutes
Voice: female-1
Narration language: English
Orientation: horizontal
Scene pacing: fast
Watermark: false
Visibility: public

Canvas: the difference is visible

The enabled Canvas video uses phrases found in the script, including “Small ideas become clear,” “Start with one question,” and “only words from this script.” The disabled video adds diagram labels that do not appear in the script: IDEA, PROCESS, OUTCOME, GROWTH, and INSIGHTS FOUND.

Sketch: a restriction, not a command

Neither short Sketch render chose to display readable words. That result is useful: turning the flag on does not force text onto every scene or require every script word to appear. It defines the permitted source if Golpo decides to generate visible text. Turning it off permits extra wording; it does not guarantee that extra wording will be used.

How to enable it in Golpo

1. Start a script workflow. The control normally appears when you select Script Only mode or Edit script before creating video.
2. Paste or approve the exact narration. Treat the script as the allowed vocabulary for generated display text.
3. Leave the checkbox selected. In the creation interface it is checked by default.
4. Generate and verify. Compare every readable word in the finished frames with the approved script, especially names, dates, claims, and regulated wording.

API v2 payload

In API workflows, set the boolean explicitly. Unlike the checked creation-interface control, the API field should not be left to an assumed default.

POST https://api.golpoai.com/api/v2/videos/generate
x-api-key: YOUR_API_KEY
Content-Type: application/json

{
  "prompt": "Create a concise visual explainer about turning complex ideas into clear stories.",
  "custom_script": "Small ideas become clear when we remove distraction. Start with one question, connect each step, and let every visual support the message. When the screen shows only words from this script, viewers can focus on the explanation.",
  "onscreen_text_from_script_only": true,
  "golpo_video_engine": "golpo_canvas",
  "canvas_style_variant": "editorial",
  "timing": "0.25",
  "narration_voice": "female-1",
  "narration_language": "english",
  "video_orientation": "horizontal",
  "scene_pacing": "fast",
  "watermark": false,
  "visibility": "public"
}

For the Sketch tests above, the only engine changes were "golpo_video_engine": "golpo_sketch" and "sketch_style_variant": "classic". The two “off” requests changed only the boolean to false.

When to keep it on—or turn it off

Frequently asked questions

Is this the same as captions?

No. The feature restricts text generated inside the visual scenes. It does not promise a full, timed accessibility caption track or transcript.

Will every word from my script appear?

No. It is an allowed-vocabulary boundary, not a requirement to display the complete narration. A scene can show one script phrase, several script words, or no readable text.

Does turning it off always create extra words?

No. The Canvas test added extra labels, while the short Sketch test did not. Turning the flag off gives the visual generator permission; the scene design still determines whether text is useful.

Does it rewrite my script?

No. This setting governs generated on-screen text. Use Script Only mode, Edit script before creating video, or API custom_script to control the narration source.