We are happy to announce that our full-stack end-to-end movie creation agent using best-in-class publicly available generative APIs is now generally available in public alpha.
Read the docs here.
This draws from and refines our previous work- Text-to-Movie in one-shot with generative agentic pipelines
You can ask for a cinematic rendition on any topic, and the agent will create a best-effort contextually accurate, full-feature film on the given topic.
Start from the movie creation API and then edit individual scenes in the Studio.
How to use?
Simply call the /generate_movie endpoint with your prompt included in the payload to get a cinematic rendition, be as descriptive as possible with a prompt up-to 500 words. This can be a great starting-point for your generative video projects. You can then edit scenes in the Studio editor.
Input Parameters
You can specify Image, Video, and Inference models for your rendition within your request payload.
When selecting image and video models, consider that each has distinct characteristics. The ideal choice depends on your specific use-case, influencing the pace and theme of your rendition.
All models perform optimally at a 16:9 aspect ratio. For 9:16 aspect ratio, Pixverse3.5 or Kling1.6 are recommended for the best results.
Each rendition passes through an agentic pipeline to determine subsequent steps, resulting in longer processing times.
We parallelize pipeline components wherever possible. However, rendering a 1-minute video typically takes over 5-10 minutes, varying by the chosen model. You can monitor progress by polling the status endpoint or setting up a WebHook URL for automatic notifications upon completion/failure.
The API is currently available to all customers on the Creators plan. Subscription-based API plans will be introduced in the future.
Other Details -
Tech Stack- Javascript stack E2E with Python and C++ embeddings.
Generative Integrations- All best-in-class publicly available APIs.
Safety - We use best-in-class OpenAI content-moderation APIs for pre-filtering prompts. Content-safety filters are further enforced on app, API Provider and Model Provider level.
Pricing - We have a single Creator's plan priced at $49.99/mo. Read more about individual model pricing and token costs here.
API Docs - Interested in integrating this API. Read the docs here.
Are you an enterprise interested in hosting the endpoint for your own custom white-labelled use-cases? Reach out at hello@samsar.one