Got Questions?
We've Got Answers.
About Visionular
Visionular builds AI-driven video compression and video intelligence technology. Its products cover video encoding (the Aurora1 AV1, Aurora4 H.264/AVC and Aurora5 H.265/HEVC encoders, plus AuroraImage for images), AI video quality and enhancement (AuroraVQA, AuroraEnhancer and the AuroraVision suite), AI video generation optimization (GenOpt) and AI video understanding (Sensa). More than 250 customers worldwide use Visionular technology.
Our unique edge lies in our AI-driven video compression technology. Our state-of-the-art Content-Adaptive Encoding (CAE) Intelligent Optimization technology integrates advanced machine learning, image processing, and image enhancement, all controlled by a quality assessment mechanism aligned with subjective standards. Our solution supports various applications, including premium VOD, live broadcast streaming, and ultra-low latency RTC.
Visionular offers three deployment models:
- AuroraSDK: integrate the Aurora encoders directly into your own pipeline (FFmpeg, GStreamer or WebRTC), on-premises or at the edge.
- AuroraCloud: a fully managed SaaS platform for VOD and live streaming.
- AuroraFlex: a self-hosted VOD and live video engine that runs on-premises or in your private cloud, so content never leaves your network.
Open-source encoders require specialized technical knowledge for optimal configuration. At Visionular, our experts finely tune the Aurora encoders to meet your specific needs, simplifying the process for you.
AI Video Encoding: Aurora Encoders and AuroraSDK
Visionular’s Aurora encoders, developed by world-renowned video codec experts, excel in delivering exceptional visual quality at lower bitrates and faster encoding speeds. Benefits include:
- Enhanced Visual Quality: Our encoders offer an improved perceptual visual experience with more detailed and aesthetically pleasing images.
- Reduced Operational Costs: Expect over 50% bitrate savings leading to significant reductions in storage and CDN expenses.
- Lower Cloud Encoding Costs: Our efficient encoders allow running multiple parallel jobs on a single machine, drastically cutting cloud encoding infrastructure costs.
Visionular offers advanced H.264/AVC, HEVC, and AV1 encoders tailored for diverse applications such as UGC, PGC, RTC, e-Learning, Sports Streaming, and Live Shopping. Our encoders are designed to outperform competitors and open-source alternatives, delivering 30-50% smaller video files, enhanced speeds, and superior video quality.
Our encoders can be utilized in multiple ways:
- SDKs: Ideal for those already using FFmpeg or GStreamer, our SDKs seamlessly integrate into your existing setup.
- AuroraCloud: This comprehensive, fully-managed SaaS platform caters to all your encoding needs for VoD, Live, and RTC scenarios.
Aurora encoders support various rate control modes like VBR, CBR, CRF, and CRF+VBV. These modes cater to a broad spectrum of applications, prioritizing either strict bitrate control or video quality. If you’re unsure about the best mode for your needs, we’re here to help with personalized advice.
AuroraSDK packages Visionular’s AI-first Aurora encoders (Aurora1 AV1, Aurora4 H.264/AVC, Aurora5 H.265/HEVC and Aurora MV-HEVC) as SDKs that drop directly into an existing pipeline, with no cloud dependency. Aurora encoders produce files 30–50% smaller than standard open-source encoders at equal visual quality.
AuroraSDK runs on x86, ARM and GPU (NVENC); integrates with FFmpeg, GStreamer and WebRTC; and supports Linux, macOS, iOS and Android.
Yes. The Aurora encoders write standards-compliant bitstreams that play back on any standards-compliant player, on TV, mobile, desktop and console, and can be packaged, protected with DRM and delivered through your existing CDN.
Our encoder pricing is based on a per-minute model, varying with the resolution of the video encoded. For instance, HD encoding is charged at double the rate of SD, and UHD at quadruple the SD rate.
AuroraCloud VOD
AuroraCloud, a SaaS platform powered by Visionular’s Aurora encoders, reduces file bitrates by 30-50% without compromising perceptual quality. It produces fully standards-compliant files playable across all devices. AuroraCloud is an investment that pays off in user retention and operational cost savings, regardless of your business model.
AuroraCloud is accessible through:
- A user-friendly web console.
- Well-documented REST-APIs for seamless workflow integration.
- AWS Triggered Events for automated operation.
Absolutely. Our REST APIs offer full functionality of AuroraCloud and come with Postman collections for easy integration. Visit our documentation for detailed guidance.
Sure! Visit our API documentation to get detailed information. Our engineers are available to discuss specific use-cases or provide deeper insights.
AuroraCloud incorporates our suite of Aurora encoders – Aurora4, Aurora5, and Aurora1 – supporting H.264/AVC, HEVC, and AV1 video coding standards.
- SDKs: Ideal for those already using FFmpeg or GStreamer, our SDKs seamlessly integrate into your existing setup.
- AuroraCloud: This comprehensive, fully-managed SaaS platform caters to all your encoding needs for VoD, Live, and RTC scenarios.
We support HLS and MPEG-DASH streaming protocols, utilizing fMP4 and TS container formats for MPEG-DASH and HLS, respectively.
Yes, we provide DRM solutions including Apple FairPlay Streaming (HLS), Google Widevine (DASH), Microsoft PlayReady, and Clear Key, with AES-128 and SAMPLE-AES encryption methods.
Definitely. AuroraCloud’s versatile template feature allows you to tailor bitrates, resolutions, and encoder settings to your requirements.
Yes, AuroraCloud is compatible with major cloud providers and easily integrates with your storage and CDN solutions.
Currently, we do not offer these services, but they are in development. Sign up for updates on this upcoming feature.
Yes, AuroraCloud enables both text and image-based watermarking. You can also overlay images and logos on your videos, specifying their position via our web console or APIs.
Pricing for AuroraCloud is based on the output minute, depending on the resolution of each output, with HD and UHD priced at multiples of the SD rate.
You can start a free trial by clicking here.
Yes – with our proprietary split-and-stitch transcoding technology, we can transcode your hour-long videos into multiple bitrates in a matter of a few minutes.
AuroraCloud is also highly scalable that allows you to submit several transcoding jobs in parallel – perfect for the highly demanding nature of the video industry.
AuroraCloud Live
AuroraCloud Live is a cloud-based live streaming platform designed for smooth, high-quality, low-latency delivery of live events to large audiences. Key features include customized ABR ladders, bandwidth reduction of 25-50%, and scalability to accommodate millions of viewers.
AuroraCloud Live is operated via our UI or our easy-to-use REST APIs, which cover a range of functionalities from stream management to metadata handling. Our APIs are well-documented and come with Postman collections for straightforward integration.
AuroraCloud Live supports RTMP, RTMPS, SRT inputs, and provides outputs in Legacy HLS, Low-Latency HLS, and WHEP formats, catering to a variety of latency requirements.
Supported codecs include H.264/AVC and H.265 for video and AAC for audio.
Yes, AuroraCloud Live is ideal for real-time, low-latency streaming, supporting both RTC and LL-HLS, including WHEP for browser-based content.
Yes, AuroraCloud Live can work with SCTE-35 markers for HLS outputs, compatible with various ad-insertion providers.
AuroraCloud Live offers live audio-to-text conversion, embedded CC passthrough, and conversion to subtitles, enhancing accessibility and viewer engagement.
Absolutely. AuroraCloud Live supports streaming with multiple audio tracks and is capable of audio-only live streaming.
Yes. AuroraCloud Live outputs RTMP(S) streams, enabling easy restreaming to platforms like YouTube, Vimeo, and Instagram.
Yes, we provide both text and image-based watermarking, as well as secure streaming through signed playback URLs using JWTs.
AuroraCloud Live pricing includes CDN delivery and encoding charges per minute of input, with additional costs based on minute of output viewed by users.
Our AuroraCloud Live API documentation is available here, providing comprehensive information on integrating and utilizing our live streaming platform.
AuroraFlex: On-Premises VOD and Live
AuroraFlex is Visionular’s self-hosted video engine for organizations that need to keep content and processing inside their own infrastructure. It is available in two editions: AuroraFlex VOD for file-based transcoding and AuroraFlex Live for 24/7 live streaming. Both run on-premises or in a private cloud.
AuroraFlex VOD brings Visionular’s AI compression engine into your own infrastructure, reducing bitrate by 30–50% while maintaining comparable or better visual quality across H.264, H.265 and AV1. It packages output in MP4, HLS and DASH, generates ABR ladders automatically, and is controlled through a full REST API with webhook callbacks.
AuroraFlex Live is a self-contained, on-premises engine that ingests, enhances and transcodes live streams with zero cloud dependency. It accepts RTMP/RTMPS, SRT and UDP/RTP/RTSP inputs up to 4K/60fps, and outputs HLS, LL-HLS, DASH, WHEP, SRT and UDP/RTP with up to 10 adaptive bitrate renditions, including AV1 and Dolby Vision.
- AuroraFlex VOD: native host installation, VM-based deployment on AWS, Azure or GCP, or containerized with Docker Compose. Storage support includes local NFS, GlusterFS and AWS S3.
- AuroraFlex Live: standard Docker containers with Kubernetes orchestration, on private cloud, public cloud (AWS, Azure, Alibaba) or bare metal.
No. AuroraFlex processes and stores everything within your own infrastructure. Outputs are written to local storage or your own bucket, under your own access controls, which supports data sovereignty and compliance requirements.
AuroraFlex Live embeds AI Video Reframing (automatic target tracking for sports), AuroraEnhancer super resolution and color enhancement, AuroraVQA no-reference quality assessment, and Frame Rate Up-Conversion. These run natively inside the engine, with no external AI service calls.
Choose AuroraFlex if you need content to stay on-premises, operate in a regulated environment, or prefer to run and scale your own infrastructure. Choose AuroraCloud if you prefer a fully managed SaaS with elastic cloud scaling and global delivery, operated by Visionular.
Request a license and installation package from your Visionular contact, deploy it with install.sh or Docker Compose, then submit transcoding tasks through the REST API or the web dashboard. Contact us to request a demo.
AI Video Quality and Enhancement: AuroraVQA, AuroraEnhancer and AuroraVision
AuroraVQA is Visionular’s no-reference (blind) video quality assessment model. It scores any video, including live streams, UGC uploads and AI-generated content, the way a human viewer would perceive it, without needing the original source clip.
PSNR, SSIM and VMAF are full-reference metrics: they compare a processed video against its original source. AuroraVQA needs no reference, so it can score quality at ingest, on live streams, or across millions of UGC uploads where the source is not available.
AuroraVQA evaluates 20 metrics across 4 quality dimensions:
- Usability: visual glitching, interlacing artifacts, sandwich effect, distortion.
- Quality damage: blur, over-sharpening, detail loss, blocking artifacts, noise, dirty lens, underexposure, overexposure.
- Aesthetics: colour, contrast, aperture.
- AIGC detection: AI-generated content detection and prompt-video consistency.
AuroraVQA achieves PLCC 0.9328 and SRCC 0.9368 on Visionular’s proprietary dataset, measuring how closely its scores correlate with human quality ratings.
AuroraEnhancer is Visionular’s AI video enhancement engine. It uses deep learning to upscale resolution up to 8K, convert SDR into HDR-like color (PseudoHDR), repair noisy or damaged footage, and increase frame rates up to 120fps.
- Smart Super Resolution (2× / 4× upscaling, up to 4K → 8K)
- Image Quality Repair (denoising, de-mosaicing, sharpening, blur removal)
- PseudoHDR
- Smart Tone Mapping (HDR to SDR)
- ROI Area Enhancement (skin tone protection, object and texture detail)
- Intelligent Frame Insertion (30/60fps up to 120fps)
Yes. AuroraEnhancer applies real-time super resolution, repair, HDR and color enhancement to live streams with latency under 1 second, for sports, esports and broadcast events.
AuroraEnhancer supports standard HDR and PseudoHDR production, tone mapping down to SDR, and Dolby Vision, HDR10, HDR10+ and HLG transcoding.
AuroraEnhancer is available as an on-device SDK, as a cloud API, and as an add-on for AuroraCloud.
Beyond encoding, video processing involves vital preparatory steps. Visionular’s AI-driven image enhancement tools optimize source content intelligently before encoding.
Our AI communicates with the encoder to consider any pre-encode enhancements, resulting in optimized video output. To explore our enhancement products, including Super Resolution, Image Quality Repair, Smart Tone Mapping, and Frame-Rate Conversion, visit our product pages.
AuroraVision is Visionular’s AI video intelligence suite, with 16+ capabilities in four categories:
- AI Video Quality: AuroraVQA, AuroraEnhancer, Frame Rate Up-Conversion, Super Resolution.
- AI Smart Media: virtual ads insertion, horizontal-to-vertical reframing, highlight generation, subtitle removal.
- AI Content Intelligence: AI video reframing, automated video clipping, video summarization, thumbnail extraction.
- AI Video Security: content moderation, Video DNA fingerprinting, invisible digital watermarking.
AuroraImage
In addition to standard JPG and PNG formats, we support advanced image coding standards like AVIF and HEIF. Visionular’s image encoders can achieve file size reductions of 30-50% compared to open-source options.
AuroraImage redefines efficiency in image compression. Deploying AuroraImage in your workflow can result in a 30-50% reduction in image file sizes, leading to:
- Substantial savings in storage costs.
- Faster image loading times.
- Increased user engagement on your platform.
- Reduced churn, driving more sales and prolonged user interaction.
53% of visits are likely to be abandoned if pages take longer than 3 seconds to load (according to Google). And, large, slow-loading images are one of the biggest culprits of slow websites.
- E-commerce, sports, and UGC websites, laden with hefty images, often fall prey to slow loading times.
- Eliminate churn and abandonment with AuroraImage – the most advanced image compression software.
- With images compressed up to 50% more efficiently than competitors, you can expect enhanced engagement, swift-loading websites, and significantly reduced abandonment rates.
AI Video Generation: GenOpt
GenOpt is Visionular’s AI-powered, pixel-level optimization engine for generative image and video workflows. It reduces the compute required for generation while improving realism, clarity, fine detail, subject consistency and temporal stability.
GenOpt delivers up to 30% token compression compared with using a generative AI model alone, at comparable visual quality. Savings vary by model and settings.
GenOpt works alongside the generative model you already use, including Google Veo, Seedance, Kling and MiniMax. You can start from a new prompt or optimize an existing AI-generated video, through an API.
GenOpt restores fine detail and sharp edges, removes the “AI plastic look” by restoring skin, facial and hair texture, and suppresses frame-to-frame flicker and visual drift through cross-frame feature matching.
AI Video Understanding: Sensa
Sensa is Visionular’s AI video understanding product. It turns raw video into structured, searchable metadata with exact start and end timestamps, so teams and viewers can search footage the way they search text.
You can search by type of scene or moment, such as action or romance scenes, and by actor or character. Sensa returns every matching segment with its start and end times.
Sensa delivers results as JSON aligned to your metadata model, ready for search and categorization.