Landscape to portrait.
No upscaling. No compromise.
Computer vision reframing that follows your subject — not the centre of the frame. Every output pixel is a direct source sample.
output
How it works
From upload to verified output in three steps
Submit a job
POST your job with the desired output ratios and content preset. Receive a presigned S3 URL for direct upload.
Pipeline processes
9 CV modules run in sequence: ingest, attention analysis, scene detection, effect classification, motion solving, EDL generation, rendering, verification, and narrative validation.
Receive verified output
A webhook fires when output is ready — or poll the API. Every output is verified for pixel provenance, sharpness, duration parity, and audio preservation before delivery.
Capabilities
Built for quality, not convenience
Most automated reframing tools silently upscale when the crop gets tight. vdowrx refuses that trade. The pipeline is built around a single invariant: if we can't produce a perfect crop, we tell you.
Crop-only guarantee
Every output pixel is a direct copy of a source pixel. No interpolation. No upscaling. No padding. A non-bypassable verifier checks every frame of every output before delivery.
Computer vision subject tracking
Faces, pose, optical flow, object detection, audio energy, and speech activity are fused into a per-frame attention vector. The crop follows what matters — not just the centre of the frame.
Multiple output ratios per job
Submit once, receive 9:16, 1:1, and 4:5 outputs from the same job. The attention pipeline runs once; the motion solver and renderer run independently per ratio.
Content-type presets
Talking head, sports, documentary, cinematic — each preset tunes the CV attention weights for its content type. A podcast preset prioritises face and speech. A sport preset prioritises motion and object tracking.
Webhooks + async delivery
Fire and forget. POST your job, upload the video, and receive a webhook when output is ready. Exponential backoff with DLQ ensures delivery. Poll the API if you prefer synchronous flow.
Quality flags, not silent failures
When output is slightly below the quality threshold but fully watchable, we deliver it with quality_flags metadata rather than failing the job. You decide what to do with the data.
Who it's for
Built for every stage of the content stack
Individual creators
Publish landscape recordings to Reels, Shorts, and TikTok — without degrading your source quality.
- ✓Talking head & podcast presets
- ✓100 jobs / 300 min per month
- ✓All output ratios: 9:16, 1:1, 4:5
- ✓URL ingestion + webhook delivery
Studios & agencies
Podcast studios, news organisations, production houses, and dev teams processing high volumes.
- ✓300 jobs / 900 min per month
- ✓10 concurrent jobs
- ✓Up to 10 team members
- ✓25 GB max file size
Media at scale
Social media teams, video agencies, and broadcast companies with high-volume or custom requirements.
- ✓Unlimited jobs & minutes
- ✓Source S3 bucket ingestion
- ✓Cloud push delivery (S3/GCS)
- ✓Custom presets + whitelabel API
Content presets
The pipeline knows what kind of video it's processing
A podcast subject barely moves. An action clip moves constantly. A nature documentary may have no clear subject at all. Content-type presets tune the computer vision attention weights so the crop behaviour matches the content — not a one-size-fits-all average.
Enterprise accounts can define custom presets with arbitrary weight vectors, saved per-account and resolved at job creation time.
See all 9 presets and their weight profiles →talking-head
High face weight, high speech weight. Smooth, centred follow for single-presenter video.
sport
High optical flow weight, high YOLO object weight. Fast, tight tracking for action content.
cinematic
Balanced weights with higher pose and flow. Respects composed shots without aggressive recentring.
nature
Low face weight, high optical flow and YOLO. Tracks animals and environment rather than human subjects.
Pricing
Simple plans. No per-output fees.
Starter
Free
120 min · 30 jobs / month
- ✓9:16 portrait output
- ✓4 starter presets
- ✓REST API access
- ✗Additional output ratios
- ✗Webhooks
Pro
$99/mo
300 min · 100 jobs / month
- ✓All output ratios: 9:16, 1:1, 4:5
- ✓All 9 presets
- ✓URL ingestion
- ✓Webhook delivery
- ✗Source bucket ingestion
Studio
$299/mo
900 min · 300 jobs / month
- ✓Everything in Pro
- ✓10 concurrent jobs
- ✓Up to 10 team members
- ✓25 GB max file size
- ✗Whitelabel API
Enterprise
Custom
Unlimited · volume pricing
- ✓Everything in Studio
- ✓Source bucket ingestion
- ✓Cloud push (S3/GCS)
- ✓Whitelabel API
- ✓Custom presets
Ready to integrate?
Start with a free account. No credit card required until you hit your first plan limit. Read the API docs, submit a test job, and verify the output quality yourself.
Create free account