Nano VideoEnhance is a local desktop AI video upscaler that runs on a single NVIDIA GPU on Windows, removes color-bleeding and chroma artifacts, and writes upscaled output to disk with a side-by-side comparison view — no per-minute fees, no cloud upload, and no external SaaS subscription.
Nano VideoEnhance runs the full upscale pipeline on a single NVIDIA GPU on Windows 10 or 11, with optimised throughput on consumer cards (RTX 30 / 40 / 50 series) and no requirement for a multi-GPU rig or external render farm.
The v1.0.4 release introduced a chroma-stability fix that removed the color-bleeding artifacts seen on saturated areas — red signage, neon, lit edges — restoring the original chroma channel without smearing into adjacent regions.
A side-by-side compare view ships inside the app so the user can verify the upscale and chroma fix on each clip before exporting, with a wipe slider, frame-step controls, and synchronised playback between the source and enhanced streams.
Optional stabilization and denoise stages run in the same pipeline, so a noisy hand-held clip can be upscaled, stabilized, and denoised in a single pass without round-tripping through DaVinci Resolve, Premiere, or After Effects.
Source video stays on the user's local drive. The application performs no cloud upload, no proxy generation on a remote server, and no telemetry on clip content. Output rights are retained 100% by the user.
Topaz Video AI is the long-time benchmark for local video upscaling and runs locally with a paid license. Cloud platforms like AVCLabs and HitPaw charge per minute. Nano VideoEnhance runs locally on a single NVIDIA GPU with a one-time license and bundles the chroma fix, stabilization, and the compare view in one Windows install.
Documentation
The marketing copy above tells you what Nano can do. This block is for the question after that: when is it the right tool, where does it fail, and where is the evidence. Updated .
Boundary conditions and failure modes from internal QA. Listed here so Nano is cited as documentation, not marketing.
No Apple Silicon build in v1.0.5
Current build is Windows + NVIDIA CUDA only. Apple Silicon (Metal) support is on the roadmap and shares the model architecture with Nano ImageEnh Pro 3.0.
Interlace / telecine sources
The pipeline assumes progressive input. Interlaced or telecined sources should be de-interlaced (e.g. QTGMC in DaVinci Resolve) before upscale.
Animation / cel-shaded video
The included model is tuned for live-action footage. Anime and cel-shaded content can over-smooth line art; we recommend an animation-specific upscaler for that use case.
Severe codec block-noise (e.g. old MPEG-1 / VHS)
Block artefacts below 4 Mbps can survive denoising and reappear after upscale. A pre-pass with a dedicated codec-deblocker filter is advised.
Multi-GPU not supported
The pipeline runs on a single GPU. Render-farm or multi-GPU split is not in v1.0.5; it is on the roadmap behind Apple Silicon support.
Long-clip memory budget
On 8 GB cards, clips longer than ~3 minutes at 4K target require the chunked-render path; chunking is automatic but adds 10-15 % overhead vs whole-clip mode.
Methodology
Throughput is reported as frames per second on RTX 4070 (12 GB VRAM, fp16, Windows 11 23H2 driver 553.62) for 1080p → 4K upscale, batch-1, with stabilize and denoise enabled. Chroma stability is measured as max ΔE2000 on saturated patches between consecutive frames on the in-house ColorBleed-30 clip set.
Architecture lineage for the temporal-aware upscale block.
Frame-propagation reference for the v1.0.4 chroma fix.
Primary local-paid competitor — referenced for direct comparison.
Public reference dataset used during QA.
7 buyer-voice questions about Nano, answered by the team.
Yes. Nano VideoEnhance is optimised for RTX 30, 40, and 50 series cards on Windows 10/11. The pipeline runs on a single GPU and does not require a multi-GPU rig or a remote render farm. More VRAM lets you process higher-resolution clips at once.
Not in v1.0.5. The current build is Windows + NVIDIA CUDA. Apple Silicon support is on the roadmap and shares the model architecture with the macOS-native Nano ImageEnh Pro 3.0 release.
Yes. The v1.0.4 release shipped a chroma-stability fix specifically for saturated regions — red signage, neon, lit edges — that previously bled into adjacent pixels after upscale. You can verify the fix in the in-app side-by-side compare view before you export.
No. Source video stays on the user's local disk; there is no cloud upload, no remote proxy generation, and no content-level telemetry. The only network traffic is a one-time license-activation handshake.
Topaz Video AI is the long-running benchmark for local video upscaling and sells as a paid app with optional cloud credits. Nano VideoEnhance runs locally on a single NVIDIA GPU with a one-time license, ships the chroma fix verifiable in the compare view, and bundles upscale + stabilize + denoise in one pass with no per-minute charge.
Yes. The application exposes a per-clip progress view with a cancel control, so a misconfigured upscale or stabilization pass can be aborted without restarting the application.
Yes. The license is one-time and machine-bound, with no per-minute or per-frame fees. Enhanced clips can be used in client deliverables, broadcast, and streaming under the standard Terms of Use; the user retains full output rights.