Bring KPainter Into Your Product
KPainter OpenAPI is for developers who want to create content inside their own product, automation, or agent workflow. It exposes three public products:
| Product | type | Best for |
|---|---|---|
| Explainer Video | video | courses, training, SOPs, research summaries, product education, and knowledge explainers |
| AI Image | image | posters, covers, illustrations, and single-image concepts |
| AI App | app | interactive pages, simulations, quizzes, and small tools |
PDFs, presentations, documents, and images are source material; they do not select the output product. Explainer Video can also use a webpage URL as source context.
Base URL
All production requests start here:
https://api.kpainter.ai/openapi/v1For example, the full capabilities endpoint is:
https://api.kpainter.ai/openapi/v1/capabilitiesA Complete Creation Flow
- Call GET /me to verify the API key.
- Call GET /capabilities to discover current options.
- Upload attachments with POST /files.
- Create one video, image, or app with POST /creations.
- Poll GET /creations/{id} for status and progress.
- Read GET /creations/{id}/outputs when delivery is available.
- Use messages to refine an Explainer Video. When a task pauses, perform only an action listed in available_actions.
Main Endpoints
| Endpoint | Purpose |
|---|---|
| GET /health | Check service availability |
| GET /me | Verify the current account |
| GET /capabilities | Discover products, parameters, templates, and models |
| POST /files | Upload input files |
| POST /creations | Create content |
| GET /creations | List the account's creations |
| GET /creations/{id} | Read status, progress, ETA, credits, and review |
| POST /creations/{id}/messages | Refine an Explainer Video |
| POST /creations/{id}/actions | Continue the latest paused task |
| GET /creations/{id}/outputs | Read QA-approved outputs |
What to Keep in Mind
- Authenticate with a Bearer API key.
- Public request and response fields use snake_case.
- Send only parameter values returned by capabilities.
- An ordinary failure does not report actual credits.
- ETA is omitted when it cannot be estimated reliably.
- Read usable results from the outputs endpoint after completion instead of constructing storage URLs.