The full-stack Text-to-Video agent is now live in public beta! You can access it via API or the VidGPT creator.
You can now create full-stack video renditions of up to 90 seconds long on any topic of your choice, with faster render times, better prompt adherence, and higher render quality. Check out this thread for 1-shot/few-shot rendition examples.
Key Features
- Effortlessly create comprehensive films and video content starting from a single text prompt. Post-process to make edits/additions.
- Provide detailed prompts with style and cinematic specs to get a high-quality movie.
- Easily post-process your video in the Studio for granular control.
- Choose from multiple best-in-class models in the Studio to find the perfect match for your project.
Updates and Improvements from Alpha
- Better prompt adherence, dialog, and visuals with a higher success rate. In our internal evaluations, over 95% of renders succeeded. (You won’t be charged for failed renders.)
- Bug fixes in Studio – Now post-process your generated movies to re-roll scenes or add/edit/update existing scenes.
What Does This Mean?
The Beta update indicates we’re moving into a stable state. We’ll also deploy fewer, more focused improvements and updates while we concentrate on delivering the best value to our customers.
Model Support
Image Models
- Imagen3
- LuminaV3
- Flux1.1Pro
Video Models
- 16:9 Aspect Ratio: Hailluo Minimax, Kling1.6, LumaRay2, Ray2Flash, Pixverse3.5, Veo2, RunwayML
- 9:16 Aspect Ratio: Kling1.6, PixVerse3.5
Chief Inference Model
- OpenAI GPT4.5
You can choose between OpenAI GPT4.5, GPTO1 and GPTO3-Mini for inference.
Chief TTS Provider
OpenAI GPT4O-Mini
Our agent infrastructure is hosted on Microsoft Azure (proud members of the Microsoft for Startups Program).
Recent Updates
- Added new models
- Enhanced success rate
- Improved retrials
- Advanced TTS models
Our client code is fully open-source. Check it out here.
We’d like to thank our API and infrastructure providers for their support: Fal (hub for model APIs), OpenAI (Inference Provider), Microsoft Azure (chief infra provider), and all other API providers. Check out all supported APIs here.
For enterprise solutions, reach out at hello@samsar.one.
We are full-stack software R&D labs based in the ASEAN region. 🤍