Pixmind
Models & tools

PixMind API / Video Tool

Video to storyboard prompts API

From reference footage to your next shot.

Turn a reference video into scene descriptions and reusable prompts. Keep the language, scene count and structured output aligned with your production workflow.

  • Read a video as scenes
  • Choose language and scene count
  • Reuse prompts in your generation workflow
Open PlaygroundAPI documentation
Reference clip from the existing PixMind video-analysis example · muted preview

Up to 30 minutes. Billed per 30 seconds, rounded up.

Try this API

Video to storyboard prompts

Use your API key to run the request. Current prices and available options load for your application.

Playground

Sign in and create an API key to run a request. TTS needs audio permission; video tools need video permission.

Upload video

No video selected

Video files up to 200 MiB

Running this request uses your API balance at the configured price.

Result

Current API credits per billing unit

Select an API key to load current prices.

Integration examples

cURL
curl 'https://aihub-admin.aimix.pro/api-platform/v1/video-to-prompt' \
  -H 'Authorization: Bearer YOUR_API_KEY' \
  -H 'Content-Type: application/json' \
  -d '{
  "videoUrl": "",
  "language": "en",
  "maxScenes": 10
}'
TypeScript
const response = await fetch("https://aihub-admin.aimix.pro/api-platform/v1/video-to-prompt", {
  method: "POST",
  headers: { Authorization: "Bearer YOUR_API_KEY", "Content-Type": "application/json" },
  body: JSON.stringify({
  "videoUrl": "",
  "language": "en",
  "maxScenes": 10
})
});
const result = await response.json();
if (!response.ok || result.code !== 1000) throw new Error(result.message);
console.log(result.data);
Python
import json, os, requests

response = requests.post("https://aihub-admin.aimix.pro/api-platform/v1/video-to-prompt",
    headers={"Authorization": "Bearer " + os.environ["PIXMIND_API_KEY"]},
    json=json.loads("{\n  \"videoUrl\": \"\",\n  \"language\": \"en\",\n  \"maxScenes\": 10\n}"), timeout=120)
response.raise_for_status()
result = response.json()
if result["code"] != 1000: raise RuntimeError(result["message"])
print(result["data"])

01 / Made for your workflow

Read a video as scenes

Analyze a reference clip and receive scene descriptions that help you organize subjects, actions and visual direction for a new storyboard.

Open Playground
Coastal road reference scene for camera-direction prompts

02 / Made for your workflow

Choose language and scene count

Select the output language and maximum number of scenes to keep the returned storyboard aligned with your working language and planning format.

Open Playground
Perfume product scene with a defined camera angle and lighting

03 / Made for your workflow

Reuse prompts in your generation workflow

Copy or download the structured result, refine its prompts, and bring them to the video model that fits your next shot.

Open Playground
Character and architectural scene for a prompt library

Use cases

Where this tool fits

Break down a reference clip

Workflow illustration

Break down a reference clip

Turn a finished sequence into scene-level notes for your creative brief and production planning.

Build a storyboard prompt library

Workflow illustration

Build a storyboard prompt library

Collect structured prompts from reference videos and adapt subjects, settings and motion for a new concept.

Connect analysis to generation

Workflow illustration

Connect analysis to generation

Use the result in an application that helps creators move from reference footage to a first set of video prompts.

Pick the tool for the output you need

Pick the tool for the output you need
WorkflowVideo to promptVideo generationVideo cleanup
Start withA reference videoText or supported reference mediaA source video
ReceiveScene analysis and promptsA new videoA processed video
Use whenPlanning shots and adapting referencesProducing footage from a promptRemoving subtitles or a fixed overlay
ControlsOutput language and maxScenesModel-specific generation parametersEngine, method and region
Next stepAdapt prompts for a video modelEdit the generated clipAdd new captions or narration

Parameters checked on October 9, 2026

API reference

From input to output

  1. 01

    Add the reference video

    Use an accessible URL or upload a local video. The analysis API accepts a duration of up to 30 minutes.

  2. 02

    Set the analysis output

    Choose language and maximum scenes, then select an API key with video permission.

  3. 03

    Save and adapt the prompts

    Wait for the task to finish, download its JSON, and adjust each prompt for your chosen generation model.

Frequently asked questions

What does video to prompt return?

It returns structured video analysis with scene descriptions and reusable generation prompts. The task result can be copied or downloaded as JSON.

Does it generate another video?

No. This route analyzes the input. To create a new video, adapt the returned prompts and submit them to a video generation API from the model directory.

How long can the reference video be?

The analysis API accepts videos up to 30 minutes. The local uploader in this playground also has a 200 MiB file-size limit.

Can I choose the prompt language?

Yes. Set the language field to the output language available in the playground. This changes the analysis output, rather than translating audio or subtitles.

What does maximum scenes control?

maxScenes sets the maximum requested scene count. It helps keep the analysis concise; it is not a request to produce that many new videos.

How is video analysis billed?

Duration is rounded up to 30-second billing units. The API-key-specific unit price is shown in the playground after selecting your key.

How do I retrieve a finished analysis?

Keep the returned task ID and query /api-platform/v1/tasks/{id} until the task finishes. Save the result JSON before passing prompts into a separate generation workflow.

Why might the input URL fail?

The server must be able to download the video. Private URLs, expired links and hosts that block server downloads can fail. Upload the local source through the media uploader when an external URL is unreliable.

Complete your workflow

Video subtitle & watermark removal

Automatically detect subtitles or erase one fixed rectangle. BASE offers inpainting or blur; PRO uses the existing advanced processing pipeline.

Text to speech

Choose a provider and voice, synthesize narration, and receive a downloadable audio URL.

Turn a reference clip into your next storyboard

Up to 30 minutes. Billed per 30 seconds, rounded up.

Open Playground