CtrlK
Topaz Labs is a software company headquartered in Dallas, Texas, that develops AI-powered tools for photo and video enhancement. Founded in 2006, the company specializes in deep learning-based solutions for image denoising, super-resolution, sharpening, and video upscaling. Topaz Labs serves over one million photographers and videographers worldwide, and its software has processed more than two billion images. The company's client base includes teams at Apple, Netflix, NASA, Nike, Google, Coca-Cola, Target, NVIDIA, and Tesla.
Topaz Labs operates primarily as a desktop software company, with its core products running locally on users' machines. This approach distinguishes it from many cloud-only AI services, as it allows photographers and filmmakers to process sensitive or proprietary visual material without uploading files to external servers. In recent years, the company has expanded into cloud-based applications and a web-based platform while maintaining its emphasis on local processing.
Topaz Labs was founded in 2006 in Dallas, Texas. In its early years, the company developed Photoshop plugins for tasks such as noise reduction, detail enhancement, and artistic effects. Products from this era included Topaz DeNoise, Topaz Detail, Topaz Clarity, Topaz Adjust, and Topaz Simplify, which were sold individually or as part of a plugin bundle. These early tools relied on traditional image processing algorithms rather than machine learning.
The company built a loyal following among professional and hobbyist photographers who used these plugins to improve image quality beyond what Photoshop's built-in tools could achieve. By the mid-2010s, Topaz Labs had established itself as one of the leading third-party plugin developers for photo editing software.
As convolutional neural networks and deep learning began to transform computer vision, Topaz Labs shifted its product strategy toward AI-powered tools. The company began training neural networks on large datasets of images to perform tasks like noise reduction, sharpening, and upscaling with results that surpassed traditional algorithmic approaches.
The first products of this new era were standalone desktop applications rather than Photoshop plugins:
These standalone AI tools represented a significant departure from the company's plugin-only heritage and marked the beginning of Topaz Labs' reputation as a leader in AI-based image enhancement.
In 2022, Topaz Labs released Topaz Photo AI, a unified application that combined the capabilities of DeNoise AI, Sharpen AI, and Gigapixel AI into a single workflow. This consolidation simplified the user experience and allowed the company to focus development resources on a single application rather than three separate products.
Topaz Video AI (originally released as Topaz Video Enhance AI in 2020, later renamed) brought the company's AI enhancement technology to video, offering capabilities like video upscaling, frame interpolation for slow motion, stabilization, and denoising.
The company has reported 1,500% revenue growth over a six-year period. As of 2025, Topaz Labs employs approximately 87 people and continues to operate from its Dallas headquarters at 5100 Belt Line Road.
In 2025, Topaz Labs expanded its product line significantly, launching cloud-based applications including Bloom (creative image upscaling), Astra (creative video upscaling), Express (browser-based quick edits), and Mosaic (photo restoration and sharing). The company also introduced Topaz Studio, a unified subscription that bundles all desktop and cloud applications.
Topaz Labs won the TIPA 2025 award for "#1 Enthusiast AI Software," recognizing its position in the consumer AI photo enhancement market.
Topaz Labs offers a portfolio of desktop applications, cloud-based tools, and mobile apps for image and video enhancement.
Topaz Photo AI is the company's flagship image enhancement application. It combines denoising, sharpening, upscaling, and face recovery into a single desktop tool. The software operates as both a standalone application and as a plugin for Adobe Photoshop, Adobe Lightroom Classic, Capture One, Apple Photos, and Affinity Photo.
| Feature | Description |
|---|---|
| Denoise | Removes noise from images, including RAW-specific denoising for files shot at high ISO settings |
| Sharpen | Restores sharpness lost to motion blur, focus miss, or general softness |
| Super Focus | Recovers detail from images with missed focus |
| Upscale | Increases image resolution by up to 16x using AI models |
| Recover Faces | Enhances facial detail in portraits, with "Realistic" and "Creative" options; doubles the resolution of previous face recovery models |
| Dust and Scratch Removal | Automatically identifies and removes physical blemishes from scanned photos |
| Adjust Lighting | Corrects exposure, highlights, and shadows |
| Balance Color | Automatic brightness and contrast correction |
| Remove | Context-aware blemish and object removal |
| Preserve Text | Recovers small details and lettering in upscaled images |
| Healing Brush | Manual tool for fixing small blemishes |
Topaz Photo AI includes 11 AI-powered tools trained on millions of images. The application supports both local and cloud rendering. Local rendering keeps all files on the user's computer, while cloud rendering offloads processing to Topaz Labs' servers for access to the latest and most powerful AI models.
The software holds a 4.8-star rating from nearly 4,900 reviews.
Topaz Video AI is a desktop application for enhancing video quality using AI models. It can sharpen footage, remove noise and compression artifacts, restore detail, stabilize shaky video, and upscale video files to higher resolutions. The software works as both a standalone application and as a plugin for Adobe After Effects and DaVinci Resolve.
Topaz Video AI includes 19 AI enhancement models, each optimized for different source material and output goals:
| Model | Purpose |
|---|---|
| Proteus | General-purpose enhancer for noise and compression artifacts |
| Iris | Restores faces and fine details in degraded footage |
| Nyx | Denoising for high-resolution input |
| Nyx XL | Preserves fine detail in high-resolution, low-light footage |
| Rhea | Specialized for 4x upscaling of textures and organic detail |
| Rhea XL | Extended version of Rhea for larger upscaling tasks |
| Artemis | Balanced model with presets for various source quality levels |
| Gaia | Upscales high-quality footage to HD/4K naturally |
| Theia | Clarity-focused sharpening and detail enhancement |
| Starlight | Diffusion model-based upscaler that transforms degraded footage to clean HD/4K |
In addition to enhancement models, Topaz Video AI includes frame interpolation models for slow motion and frame rate conversion:
| Model | Purpose |
|---|---|
| Apollo | Smooth slow motion and frame rate conversion |
| Apollo Fast | Faster variant of Apollo for quicker processing |
| Chronos | Handles challenging frame rate conversions with precision |
| Chronos Fast | Faster variant of Chronos |
| Aion | Built for high-resolution inputs and extreme slow motion |
Other video processing features include motion deblur, stabilization, grain addition, and SDR to HDR conversion. The software supports both local and cloud rendering.
Topaz Video AI holds a 4.8-star rating from over 3,500 reviews.
Topaz Gigapixel AI is a dedicated image upscaling application. It enlarges images by up to 6x resolution using AI models that create new pixels with realistic detail rather than relying on simple interpolation. The software works as both a standalone application and as a plugin for Photoshop and Lightroom Classic.
Gigapixel AI includes nine specialized enhancement models:
| Model | Best For |
|---|---|
| Standard | General-purpose balanced enhancement |
| High Fidelity | High-resolution image processing |
| Low Res | Small, heavily compressed files |
| Text & Shapes | Scanned documents and old photos with text |
| Art & CG | Digital artwork and illustrations |
| Recover | Detail restoration from degraded images |
| Redefine | Reconstruction of heavily damaged photos |
| Face Recovery | Portrait-specific facial enhancement |
| Wonder Model | All-in-one upscaling with one-click processing |
Topaz Gigapixel AI holds a 4.8-star rating from over 5,400 reviews, making it the most-reviewed product in the Topaz Labs lineup.
Bloom is a web-based creative upscaling tool designed for AI-generated images and digital artwork. It can upscale images by up to 8x resolution with outputs up to 100 megapixels on the Pro plan. Unlike the company's desktop tools, which focus on faithful detail preservation, Bloom takes a creative approach to upscaling by adding contextually relevant detail and texture guided by user prompts.
Key features of Bloom include adjustable creativity levels, prompt influence for guiding the upscaling process, a "Realism Mode" that adds natural-looking texture, and a "Wonder 2 Mode" that prioritizes preservation of existing detail. Users can generate four unique variations of each upscale simultaneously.
Bloom offers a free tier with 10 images per month, making it one of the few Topaz Labs products with a no-cost option.
Astra is a web-based creative video upscaling tool, primarily designed for AI-generated video content. It is powered by the Starlight AI model family and can upscale videos to 4K resolution. Astra offers two upscaling modes: Precise Mode (faithful to the original look) and Creative Mode (with "Subtle" and "Bold" intensity options). The tool also includes frame interpolation for adjusting frame rates up to 120fps and creating slow-motion effects up to 8x.
Astra operates on a credit-based system, and credits roll over indefinitely. Adobe integrated Astra as a partner model in its Adobe Firefly platform in December 2025, expanding Topaz Labs' reach into Adobe's ecosystem.
Express is a browser-based application that brings several of Topaz Labs' core AI capabilities to the web without requiring desktop software installation. Features include denoising, unblur (sharpening), face recovery, lighting adjustment, image sharpening, image upscaling (up to 4K), and video upscaling (up to 4K). Express also includes an experimental "Starlight" feature for enhancing degraded, low-resolution, or aged video.
Mosaic is a cloud-based photo restoration and sharing application designed for digitizing and restoring old photographs. Users photograph physical prints with their phone, and the application automatically corrects for glare and angles. AI-powered restoration features include dust and scratch removal, colorization of black-and-white photos, and noise reduction. Mosaic also provides organizational tools for creating albums and sharing restored photos with family members. As of early 2026, Mosaic is listed as "Coming soon" with a waitlist.
Gigapixel iOS brings the company's image upscaling technology to Apple mobile devices. It allows users to upscale images directly from their iPhone or iPad.
Topaz Labs develops proprietary AI models for image and video enhancement. The company's approach relies on training deep neural networks on large datasets of image and video pairs. For denoising, models are trained on pairs of noisy and clean images, learning to distinguish between genuine image detail and noise artifacts. For upscaling, models learn the relationship between low-resolution and high-resolution versions of the same image, enabling them to generate plausible high-frequency detail when enlarging photos.
The company's core technical capabilities span three domains:
Starlight is Topaz Labs' family of generative AI models for video enhancement. The company describes Starlight as the "world's first generative video upscaling model." Unlike traditional upscaling approaches that work frame by frame, Starlight uses diffusion model architectures to transform degraded footage into clean HD or 4K video. Starlight Mini is described as the "first local diffusion model," capable of running on consumer hardware without requiring cloud processing.
The Starlight models power both the desktop Topaz Video AI application and the cloud-based Astra service.
The Wonder Model is a unified AI model that combines upscaling, sharpening, and denoising into a single operation. Rather than applying these enhancements sequentially (which can compound artifacts), the Wonder Model processes all three tasks simultaneously. The Standard MAX Model, a related model, is described as faster than first-generation diffusion models while maintaining high output quality.
The Wonder Model is available in both Topaz Photo AI and Topaz Gigapixel AI. On the desktop, it requires an NVIDIA GPU with at least 8GB of VRAM. Cloud rendering makes the model accessible to users with less powerful hardware.
Topaz Labs offers two processing modes across its product line:
| Processing Mode | Advantages | Requirements |
|---|---|---|
| Local Rendering | Privacy (files stay on user's computer), no internet required, no per-image costs | Capable GPU (6GB+ VRAM), 16GB+ RAM, 50-60GB storage |
| Cloud Rendering | Access to latest/most powerful models, no hardware limitations, faster on underpowered machines | Internet connection, cloud credits or subscription |
The local processing option is a significant differentiator for Topaz Labs. Many professional photographers work with client images that are subject to confidentiality agreements, and the ability to enhance these images without uploading them to cloud servers is a practical requirement for these workflows.
| Platform | Minimum Requirements |
|---|---|
| Windows | Windows 10 or 11, Intel/AMD CPU with AVX instructions (post-2016), 16GB RAM, 6GB VRAM (NVIDIA, AMD, or Intel ARC), 60GB storage on C: drive |
| macOS (Apple Silicon) | macOS 13 or higher, Apple M-series chip, 16GB RAM, 50GB storage |
| macOS (Intel) | macOS 13+, Intel 2016+ CPU, 16GB RAM, 4GB VRAM (8GB recommended); not recommended due to slower performance |
Generative AI models (such as Starlight and Wonder) require macOS 14 or higher and are significantly slower on Intel Macs. The Wonder 2 local model requires an NVIDIA GPU with 8GB or more of VRAM. Virtual machines, NAS setups, external GPUs, and Linux are not officially supported.
Topaz Labs uses a subscription pricing model with individual product plans and a bundled option called Topaz Studio. Pricing is structured into Personal and Pro tiers.
| Product | Personal (Annual) | Personal (Monthly) | Pro (Annual) | Pro (Monthly) |
|---|---|---|---|---|
| Topaz Photo AI | $17/mo ($199/yr) | $21/mo | $50/mo ($599/yr) | $58/mo |
| Topaz Video AI | $25/mo ($299/yr) | $33/mo | $58/mo ($699/yr) | $67/mo |
| Topaz Gigapixel AI | $12/mo ($149/yr) | $17/mo | $42/mo ($499/yr) | $50/mo |
| Tier | Annual | Monthly |
|---|---|---|
| Studio Personal | $33/mo ($399/yr) | $69/mo |
| Studio Pro | $67/mo ($799/yr) | $75/mo |
Topaz Studio includes all desktop applications (Photo, Video, Gigapixel) and all cloud applications (Bloom, Astra, Express, Mosaic) with unlimited local rendering, unlimited cloud image rendering, and 300 to 600 monthly video cloud credits depending on the tier.
| Application | Personal (Annual) | Personal (Monthly) |
|---|---|---|
| Bloom | $19/mo ($229/yr) | $39/mo |
| Astra | $39-$299/mo | Varies |
| Express | $12/mo ($140/yr) | $25/mo |
| Mosaic | $3.75/mo ($45/yr) | $9/mo |
| Gigapixel iOS | $12/mo ($140/yr) | $15/mo |
Personal plans allow limited commercial use for organizations with under $1 million in annual revenue. Pro plans include full commercial use rights, additional cloud concurrency, and access to the most powerful local AI models.
Topaz Labs also offers enterprise pricing with custom plans available through its sales team.
Topaz Labs' tools are widely used in professional photography workflows. Common applications include:
Topaz Video AI is used in film and video post-production for:
Archivists and preservationists use Topaz Labs tools to restore and enhance historical photographs and film:
With the rise of AI image generation tools like Midjourney, DALL-E, and Stable Diffusion, Topaz Labs has found a growing user base among AI artists. These creators use Bloom and Gigapixel AI to upscale AI-generated images, which are often produced at relatively low resolutions, to print-quality or display-quality sizes with added detail and texture.
In late 2025, Topaz Labs became a partner in Adobe's Firefly ecosystem. Adobe integrated Topaz Astra as a partner model within the Adobe Firefly platform, making Topaz Labs' video upscaling technology available to Firefly users. This partnership was announced alongside other partner integrations at Adobe MAX 2025 in October and was expanded in December 2025 when Astra-powered video upscaling capabilities were added to the Firefly platform.
This integration marks a notable expansion for Topaz Labs beyond its standalone product offerings and into the broader creative software ecosystem.
Topaz Labs competes with a range of AI-powered image and video enhancement tools.
| Competitor | Type | Key Strengths |
|---|---|---|
| Adobe Enhance (in Photoshop/Lightroom) | Built-in tool | Integrated into existing Adobe workflows; no additional cost for Creative Cloud subscribers |
| ON1 NoNoise AI | Desktop app | Combined noise reduction and sharpening; part of ON1 Photo RAW suite |
| DxO PureRAW / DeepPRIME | Desktop app | Specialized RAW denoising using DxO's camera/lens profile database |
| Luminar Neo (Skylum) | Desktop app | AI-powered photo editing with a broader feature set beyond enhancement |
| NVIDIA RTX Video Super Resolution | GPU feature | Real-time video upscaling built into NVIDIA drivers; limited to RTX GPUs |
| Video2X | Open-source | Free, open-source video upscaling using neural networks |
| BigJPG | Web service | Cloud-based image upscaling using deep learning |
| Pixelmator Pro | macOS app | Apple-native photo editing with ML-powered enhancement features |
Topaz Labs differentiates itself through its breadth of AI models (19 video models, 9 upscaling models, 11 photo enhancement tools), its emphasis on local processing for privacy, and its focus on professional-grade output quality. The company's 1,500% revenue growth over six years suggests strong market demand for its approach.