Topaz Labs has launched a new REST API that provides third-party developers, streaming platforms, and enterprise workflows with access to its proprietary AI-based video and image upscaling models. The service includes tiered subscription options and supports on-premise deployment for enterprise clients.
This release shifts Topaz from a prosumer desktop tool to a scalable infrastructure component for streaming platforms and production houses. By exposing models like Astra and Starlight via API, Topaz provides a direct path for automate-at-scale enhancement of massive libraries or user-generated content. For the broader ecosystem, it intensifies competition with real-time solutions like NVIDIA Maxine by prioritizing archival-quality output over sub-500ms latency. Executives should monitor how this affects the 'ingest at 1080p, deliver at 4K' cloud-rendering pattern, particularly as enterprise clients seek to avoid hardware bottlenecks through Topaz’s new on-premise and cloud credit models.
Topaz Labs has built its API offering around three model families: Precision, Generative, and Creative, each addressing different enhancement needs for enterprise workflows. The company's developer documentation confirms that the API provides video models including Proteus for precision upscaling, Starlight for generative upscaling, and Astra for creative upscaling, alongside frame interpolation and compression recovery capabilities. The API enforces a 500MB request size limit and uses HTTPS-only communication, with rate limiting that returns HTTP 429 responses during high server load. Topaz positions these models as requiring no manual tuning, with default automatic settings designed to produce the highest quality outcome at scale, a deliberate departure from the interactive desktop workflow that previously defined the company's product line.
The company claims its enhancement models are used by more than 3 million users worldwide and trusted by teams at Google, Tesla, Nike, and Nvidia as a last-mile solution for improving image and video quality. All models are developed in-house by PhD researchers at Topaz's Dallas, Texas headquarters. The API uses a credit-based pricing system where one credit processes a single request at up to 24 megapixel output size, equivalent to a 6000 by 4000 pixel image through the Enhance endpoint. Topaz offers both cloud-based processing and on-premise deployment for enterprise clients with restricted workflows, and the service requires no credit card to begin testing.
Astra 2 focuses on transforming AI-generated and stylized video with bold new detail and prompt-guided direction. The API's video endpoint accepts source metadata including resolution, container format, duration, frame rate, and frame count, then returns a download link once processing completes. This architecture positions Topaz as a batch-processing enhancement layer rather than a real-time pipeline component, distinguishing it from GPU-accelerated real-time solutions that prioritize latency over output fidelity.
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source