# Zorq AI

> Zorq AI is an AI image and video generation platform for creators, marketers, and teams.

Canonical: https://www.zorqai.io
Primary-Language: en

## What Is Zorq AI

Zorq AI is a browser-based platform for generating AI video and images. It runs several models side by side under one credit balance — Grok Imagine, Veo 3.1, Seedance 2 and Kling for video, Nano Banana 2 and GPT Image 2 for stills — so a still can be generated and then animated without leaving the site. Nothing to install and no GPU required.

Common tasks it is used for: turning a photo into a short video, animating a still by following a reference motion clip, making a photo talk from a typed script, generating stills for concepts and product frames, and turning a selfie into a collectible figure.

## AI Models

### Video Generation
- **Grok Imagine** — The default and free-tier video model. Text-to-video and image-to-video, 6-15 second clips, 480p to 1080p. 2 credits/sec at 480p, 3 at 720p, 6 at 1080p; free accounts get 480p and 720p at up to 6 seconds. Runs on the free signup credits.
- **Veo 3.1** — Paid. Text-to-video and image-to-video, priced flat per clip: Lite ~24 credits, Fast ~44, Standard ~70 (1080p). Model page: /veo-3-1
- **Seedance 2** — Cinematic text-to-video and image-to-video. Mini (480p) ~6 cr/s, Fast (720p) ~21 cr/s, Standard (720p) ~26 cr/s, Pro (1080p) ~59 cr/s. Paid only. Model page: /seedance-2
- **Kling O3** — Text/image-to-video with multi-reference support (up to 7 images). Std ~8 cr/s, Pro ~11 cr/s (with audio ~13/16). Model page: /kling-o3
- **Kling O3 Edit** — Video-to-video editing. Upload a source video and describe the edit in a prompt.
- **Kling V3 Motion Control** — Upload a start image + a motion reference clip to animate a still photo. Std ~7 credits/sec (720p), Pro ~12 credits/sec (1080p). Paid. Model page: /kling-motion-control
- **Kling V2.6 Motion Control** — Same workflow, previous generation. Free-tier usable.
- **Kling Avatar 2.0** — Audio-driven talking photo. One photo + a script; the voice-over is generated automatically (ElevenLabs TTS) and the face lip-syncs to it. Standard 5 credits/sec of speech, Pro 10. Speech 2–60s; measured ~3–4 minutes to render a 10-second clip; output 960x960 with audio muxed in. Tool page: /ai-talking-photo

### Image Generation
- **Nano Banana 2** — The default image model for stills, concepts and product frames. Model page: /nano-banana-2
- **GPT Image 2** — Higher-detail image generation; quality x resolution pricing (low/medium/high at 1K/2K/4K). Model page: /gpt-image-2
- Image workspace: /image. Ready-made preset: /ai-action-figure-generator (photo to collectible figure).

## Pricing

### Subscription Plans (Monthly)
- **Lite** — $15.9/mo, 600 credits/mo
- **Pro** — $35.9/mo, 1620 credits/mo (most popular)
- **Premium** — $69.9/mo, 3495 credits/mo

### Subscription Plans (Yearly — save ~17%)
- **Lite** — $159/yr, 7200 credits/yr
- **Pro** — $359/yr, 19440 credits/yr
- **Premium** — $699/yr, 41940 credits/yr

### One-Time Credit Packs
- **Mini Pack** — $6.9, 170 credits
- **Starter Pack** — $19.9, 540 credits
- **Creator Pack** — $39.9, 1190 credits
- **Studio Pack** — $69.9, 2690 credits

### Free Tier
- New users receive 50 free credits on sign-up, no card required — enough for 2 Grok Imagine clips at 720p.
- Free tier covers Grok Imagine (480p/720p, up to 6s) and Kling V2.6 Motion Control. Other models, including Veo 3.1, require credits.

## Key Features
- Motion control video generation (animate any image with a reference motion clip)
- Motion library with preset clips (dance, walk, gesture, etc.)
- Text-to-video and image-to-video generation
- Video editing with AI (restyle existing footage)
- Start/end frame control for precise video output
- Audio generation with video
- Pro mode for higher quality 1080p output
- No watermark on paid plans
- Private generation on paid plans
- Priority queue on paid plans

## Public Pages
- / : Home page for Zorq AI
- /video : AI video generation — motion control, text-to-video, video editing. Includes examples, motion library, and FAQ.
- /image : AI image generation workspace with examples and FAQ.
- /kling-motion-control : Kling Motion Control model page — image-to-video with motion presets, specs, pricing, FAQ.
- /kling-video-generator : Kling AI Video Generator page — Kling video models, workflow, pricing, FAQ.
- /ai-image-to-video : AI image to video landing page — workflow explanation and examples.
- /ai-action-figure-generator : AI action figure generator — turn one photo into a collectible figure (boxed vinyl, desk-scale, or chibi). Runs on the image models; generic unbranded figures only.
- /seedance-2-5 : Seedance 2.5 model page — ByteDance Seedance 2.5 running live on Zorq AI, text-to-video and image-to-video. Its defining capability is length: ONE continuous take of 4-30 seconds, the longest on the site (Veo 3.1 stops at 8s, Kling O3 at 15s). Two tiers on our provider, 480p at 19 credits/sec and 720p at 41 — no 1080p or 4K tier here. Optional generated audio; 16:9 / 9:16 / 1:1; MP4 out.
- /ai-photo-restoration : AI photo restoration — repair creases, tears, stains, dust and fading in one pass on an uploaded scan. Runs on Nano Banana 2; the exact restoration prompt is published on the page. ~50s measured.
- /ai-talking-photo : AI talking photo generator — type a script, it is voiced automatically, and the face in your photo lip-syncs to it (Kling Avatar 2.0). ~4 min per clip; 5 credits/sec standard, voice included.
- /ai-video-upscaler : AI video upscaler — take a low-resolution or compressed clip to 1080p or 4K at 2x/4x. Runs Topaz Video Upscaler server-side, billed per second of source (6 credits/sec at 2x, 10 at 4x) rather than as a ~$299 desktop licence. Clips ≤5s covered by signup credits; 32MB, MP4/MOV/MKV, 1-60s. Does not deblur, deshake or remove watermarks.
- /ai-clothes-changer : AI clothes changer — describe an outfit in words and it replaces the clothing in one uploaded photo while the face, pose, body shape and background stay put. Runs on Nano Banana 2 (GPT Image 2 optional); the exact prompt is published on the page. ~30s measured, one photo per run. Describe-only: it does not composite a specific real garment, cannot confirm sizing, and does not do video.
- /ai-hairstyle-changer : AI hairstyle changer — describe a haircut or a hair colour and it rebuilds the hair in one uploaded photo while the face, skin tone, clothing and background stay put. Works the same for men's cuts and for colour-only changes (highlights, roots, balayage) with no separate mode. Runs on Nano Banana 2 (GPT Image 2 optional); the exact prompt is published on the page. 12-30s measured, one photo per run. It cannot account for your hair's density or curl pattern, and does not do video.
- /ai-aging-filter : Aging filter / AI age progression — ages (or de-ages) the face in one uploaded photo while keeping the same person, clothing and background. Runs on Nano Banana 2 (GPT Image 2 optional); the exact prompt is published. MEASURED CALIBRATION, published on the page with all five source images: the model under-shoots the requested age by 15-25 years (asked 40 → reads ~30; asked 60 → ~45; asked 75 → ~50; asked 80 → ~65; asked 95 → ~70) and tops out around 70 no matter how high you ask; de-ageing under-shoots the same way (asked 18 → reads ~24). So ask for 15-25 years past the age you want. Entertainment only — it is not a prediction of how someone will age, and must not be used for forensic or missing-person age progression. Stills only, no video.
- /photo-to-anime : Photo to anime converter — redraws one uploaded photo as an anime/illustration while the person stays recognisable (hair, glasses, clothing, pose, expression carry across). Runs on Nano Banana 2 (GPT Image 2 optional); the exact prompt is published. MEASURED, published with all four runs on one portrait: asking for 'modern anime with large expressive eyes' returns a NEAR-PHOTOREALISTIC illustration, while naming a sub-genre ('soft shoujo') produces real anime eyes on the first try — name the sub-genre, not the adjective. The style also decides whether the background survives: the modern pass flattened a lit wall to plain colour; the 90s-cel and ink-watercolour passes kept the window frame and its shadow. Deliberately does NOT offer named studio styles (trademarked, and the studios have objected to AI imitation) — describe line weight, palette and shading instead. Stills only, one photo per run, no video.
- /ai-family-photo-generator : AI family photo generator — merges separate portraits into one family picture, matching scale and lighting across every face, for the case where the family was never in the same room (a grandparent who has passed, a sibling overseas, a baby born after the last reunion). Runs on the image models (nano-banana-edit / seedream edit); the exact prompt is published. 3-5 people per frame is the reliable band; past about six, faces at the back soften and merge. HONEST LIMIT stated on the page: the five published images are style runs from our pipeline, NOT a published side-by-side of several uploads being merged — we did not stage a fake before/after. Stills only.
- /anime-character-maker : Anime character maker — creates an ORIGINAL character from a written description and returns a design sheet (front, three-quarter and back views plus expression studies), rather than assembling a face from preset parts the way picrew-style creators do. Runs on the image models; the exact prompt is published. The clause that does the work is 'consistent hair, eye colour, outfit and proportions across every panel' — without it you get three attractive characters who are not the same person. Reuse the written description, not the image, to bring a character back. Concedes openly that preset dress-up creators beat this on exactness and reproducibility. Hands, held props and lettering remain unreliable. Distinct from /photo-to-anime, which redraws an existing photograph.
- /ai-combine-two-images : Combine two images with AI — merges two uploaded pictures into one frame: a person into a scene, two portraits side by side, or a double exposure where one picture becomes the texture filling the other's silhouette. Runs on nano-banana-edit. Name which image is the subject and which is the setting, or the model reassigns the roles differently on every run — that is the single biggest cause of perceived inconsistency. 'Cohesive colour grade across the whole frame' is what removes the visible seam. Two inputs is the reliable maximum; for groups use /ai-family-photo-generator. This is a generative redraw, not a layer stack: the same pair produces a different composition each run, and all four of our published passes differ to show that.
- /sketch-to-video-ai : Sketch to video AI — animates a drawing you supply instead of re-rendering it into a different style. Runs on the video generator (reference-to-video / Kling image-to-video). The decisive prompt clause is 'no re-rendering into a different art style' alongside the preservation wording; dropping it is the usual reason people say the tool ruined their drawing. Inked line art with flat colour survives well; faint graphite sketches confuse it, because thin grey lines read as texture rather than edges — raise contrast or ink first. Direct ONE movement, not a story. 12-15s clips; art drifts further from the source the longer the clip. HONEST LIMIT: both published examples started from finished illustrated artwork, not pencil roughs.
- /how-to-add-a-person-to-a-photo : Guide — three tested ways to add a missing person to an existing photo, compared by how they fail. (1) AI generator: relights everyone to one light source, so it is the only method that survives mismatched lighting, but it redraws the whole frame. (2) Phone cut-out: about a minute, moves pixels without relighting, so the join shows when the two photos disagree about the sun. (3) Desktop editor with layers: the only method that leaves the original literally unchanged, and about thirty minutes of work. The two prompt clauses that decide the AI result are the position ('at the same distance from the camera as the other people' — omitting it is why added people come out subtly the wrong size) and the lighting match. Funnels to /ai-combine-two-images.
- /how-to-turn-a-photo-into-a-line-drawing : Guide — converting a photograph to line art, with five published runs. The real decision is edge filter vs generative model, and they produce different things: a filter traces what is literally there (accurate, free, reproducible, and it traces noise as line work), while a generative model recognises the subject and decides which lines carry meaning (varied line weight, deliberate blank space, invents detail, never identical twice). 'No grey wash, no shading gradients' is the single most common fix when output comes back looking like a grey photo. HONEST STEER AGAINST OURSELVES: for buildings, products and technical subjects an edge filter is more accurate and free — use that instead. Portraits are the best case; groups hold up to about five faces.
- /grok-imagine-alternative : Comparison — six alternatives to Grok Imagine for anyone who does not want an X Premium subscription. FACT CORRECTION, because the SERP is full of the opposite claim: Grok Imagine is NOT discontinued. What users widely reported (around 19 March 2026, with no xAI announcement) is the removal of the FREE tier, with access running through X Premium since — the page attributes that date to community reporting rather than presenting it as official, and contrasts it with Sora, which IS formally discontinued. Lists Veo, Kling, Runway, Hailuo and local open-source models alongside this site. States plainly that anyone leaving over content filtering will be MORE restricted on every hosted option including ours; local models are the only honest answer there.
- /sora-alternatives : Comparison — seven Sora replacements, published with the dates from OpenAI's own help-centre article 'What to know about the Sora discontinuation'. PRIMARY SOURCE, quoted: 'The Sora web and app experiences were discontinued on April 26, 2026.' and 'The Sora API will be discontinued on September 24, 2026.' — the second date is the migration deadline for anyone who built on the API, and most alternatives round-ups omit it. OpenAI also states that after any final export window passes it will 'permanently delete any data associated with your use of Sora'; export runs from sora.chatgpt.com/sunset, and purchased ChatGPT/Sora credits can still be used for Codex. The page's advice is to export BEFORE choosing a replacement. Options compared: this site, Veo via the Gemini API, Kling, Runway, Luma, Hailuo, and local open-source models.
- /veo-3-free-alternative : Comparison — what 'free' actually buys you across five Veo 3 alternatives, converted into clip counts rather than adjectives. CLAIMS BOUNDARY, stated on the page: the signup bonus here is 50 credits, which is about TWO clips on Grok Imagine, the free-tier model — a trial, not a free tier, and never described as free Veo or unlimited; Veo 3.1 itself is paid here. Names Google's own Gemini free allowance as option one, i.e. real Veo ahead of our substitute. Four kinds of 'free' distinguished: signup bonus (finite), daily allowance (Kling, genuinely recurring), watermarked tier (unusable output), local models (the only unconditional free). ARGUES AGAINST SUBSTITUTING when the job needs synchronised audio (Veo 3.1 has it, most alternatives do not), real physics (water, cloth, crowds, people walking), or a deadline. No unlimited Veo tier exists at any price.
- /ai-birthday-photos : AI birthday photos — builds a milestone birthday shoot around one uploaded photo: the marquee number, the flower wall, the balloon arch, the staged corner a photographer would spend two hours dressing. Runs on the image models; the exact prompt is published. Name the PROP, not the occasion ('a large illuminated marquee number four'), and state the age twice — as a word and as a numeral — because models drift children toward a generic mid-childhood face. Keeping one prompt template and changing only the numeral produces a consistent multi-age set in one session. One subject per frame; for the whole family use /ai-family-photo-generator. MEASURED AND PUBLISHED: lettering is a coin flip — sometimes clean, sometimes a plausible near-miss with one letter wrong in a name — so generate without text and set the type yourself before printing.
- /pet-dance-ai : Pet dance AI — turns one photo of a dog or cat into a short dancing clip. Image-to-video on the video models; no lip-sync gimmick, the pet keeps its own markings and body shape.
- /pricing : All plans and credit packs with comparison
- /library : Motion reference clip library for browsing presets
- /blog : Blog index and articles
- /about : About page
- /privacy-policy : Privacy policy
- /terms-of-service : Terms of service
- /refund-policy : Refund policy

## German Pages (standalone, not translations)
- /de/bild-zu-video-ki : Bild zu Video KI — ein Foto hochladen und einen kurzen Clip mit natürlicher Bewegung erhalten.
- /de/alte-fotos-restaurieren : Alte Fotos restaurieren — Risse, Knicke, Flecken und Verblassen in einem Durchgang reparieren.
- /de/pricing : Preise und Credit-Pakete.

## Blog Articles
- /blog/ai-video-motion-control-workflow-sop : AI Video Motion Control Workflow: A 30-Minute SOP (Zorq AI) — Learn a practical 30-minute motion control SOP: plan shots, pick a model, iterate fast, and review versions in Zorq AI with a click-by-click checklist.
- /blog/creative-brief-matrix-motion-control-templates : Creative Brief Matrix for Motion Control (4 Templates + Zorq AI Setup) — Use a 2x2 creative brief matrix (awareness vs conversion × safe vs bold) to plan Motion Control shots. Includes 4 templates, review gates, and Zorq AI steps.
- /blog/motion-control-model-picker-kling-v3-v2-6-nano-banana-2 : Model Picker for Motion Control: Kling v3, v2.6, Nano Banana 2 — A practical motion control model picker: choose Kling v3, Kling v2.6, or Nano Banana 2 based on your goal, assets, and approval loop—inside Zorq AI.
- /blog/motion-control-qa-checklist-2026 : Motion Control QA Checklist: Catch Drift Before Client Review — A practical motion control QA checklist for AI video teams: stop identity drift, flicker, and off-brand frames before you share drafts for approval.
- /blog/motion-control-start-frame-lock-complete-guide : Start Frame Lock: The Complete Guide for Motion Control (No Drift) — Start frame drift ruins motion control. This complete guide shows how to lock a start frame, run one-change iterations, and pass a review gate without rework.
- /blog/zorq-ai-review-workflow-first : Zorq AI Review (2026): A Workflow-First Look at Motion Control — A workflow-first Zorq AI review: who it’s for, how motion control fits a still→review loop, what to test in 30 minutes, and when to pass.
- /blog/start-frame-style-guide-template : Start Frame Style Guide Template for AI Video Teams — A start frame style guide template to keep AI video drafts consistent across iterations, reduce review time, and prevent subject drift in motion control workflows.
- /blog/start-frame-approval-checklist : Start Frame Approval Checklist: Lock Consistency Before Motion — A practical start frame approval checklist for AI video teams: lock composition, brand rules, and reject criteria before running motion control.
- /blog/sora-shuts-down-creative-team-workflow : AI Video Workflow Continuity: How Creative Teams Reduce Model Risk — A practical continuity plan for AI video teams: reduce dependence on one model, keep still-to-motion assets reusable, and protect review workflows.
- /blog/nano-banana-2-vs-nano-banana-pro : Nano Banana 2 vs Nano Banana Pro: Are They the Same? — Nano Banana 2 vs Nano Banana Pro explained: which is newer, did Nano Banana 2 replace Pro, what changed, and how to use Nano Banana 2 from 7 credits on Zorq AI.
- /blog/nano-banana-2-creative-brief-workflow : Nano Banana 2 Creative Brief Workflow (From Idea to First Draft) — A practical creative-brief workflow for Nano Banana 2: define the shot, pick a direction, generate a still, then add motion control with a review rubric.
- /blog/motion-control-review-checklist : Motion Control Review Checklist: Approve AI Video Drafts Faster — A motion control checklist for marketing teams: reduce revisions, keep subjects consistent, and approve AI video drafts faster with a repeatable review loop.
- /blog/kling-v3-vs-v26-motion-control : Kling v3 vs Kling v2.6 Motion Control: Which to Pick? — Kling v3 vs Kling v2.6 motion control—use this decision guide to choose the right model for your brief, reduce revisions, and approve AI video drafts faster.
- /blog/kling-v3-vs-v26-motion-control-what-changes : Kling v3 vs Kling v2.6 Motion Control: What Changes in Practice — A practical comparison of Kling v3 Motion Control vs Kling v2.6 Motion Control: iteration speed, stability, control, and how teams should choose per shot.
- /blog/kling-v3-vs-kling-v26-motion-control-for-teams : Kling v3 vs Kling v2.6 Motion Control: Which One Should Your Team Use? — A practical team guide to choosing between Kling v3 and Kling v2.6 Motion Control by stage, speed, and review requirements.
- /blog/kling-motion-control-aspect-ratio-guide : Kling Motion Control Aspect Ratio Guide: 9:16 vs 1:1 vs 16:9 — How to choose the right aspect ratio for Kling motion control (9:16 vs 1:1 vs 16:9), what to lock first, and how to keep iterations comparable.
- /blog/is-zorq-ai-safe-trust-checklist : Is Zorq AI Safe? A Practical Trust Checklist — A practical checklist to evaluate whether Zorq AI is safe for your team: what to review on privacy/terms, how to reduce risk, and how to start small.
- /blog/faceswap-ai-motion-control-workflow : FaceSwap AI for Product Videos: A Safe, Repeatable Workflow — A practical FaceSwap AI workflow for product videos: start with a still, control motion, run a review gate, and avoid common privacy and consent mistakes.
- /blog/best-sora-alternatives-2026 : Sora Alternatives in 2026: A Workflow-Based Evaluation Guide — Evaluate Sora alternatives in 2026 with a practical workflow checklist: control, still-to-motion fit, review speed, and production consistency.
- /blog/best-ai-video-tools-after-sora : AI Video Workflow Tools After Sora: A Practical Evaluation Guide — A workflow-first evaluation guide for AI video tools after Sora: compare control, still-to-motion fit, iteration speed, and team review needs.
- /blog/asset-pack-script-template : 30-Second AI Video Ad Script Template (Asset-Pack Workflow) — A practical 30-second AI video ad script template that pairs with an asset-pack workflow: faster reviews, consistent start frames, and fewer motion-control reruns.
- /blog/ai-video-motion-control-troubleshooting-guide : AI Video Motion Control Troubleshooting: 12 Failures + Fixes — Fix motion control outputs faster with a practical diagnostic loop and 12 common failure modes—drift, warping, flicker, pacing, and more.
- /blog/ai-video-iteration-log-motion-control : The AI Video Iteration Log: A Simple System for Motion-Control Teams — Stop guessing why a version worked. Use an iteration log to run motion-control experiments, speed approvals, and make AI video output repeatable.
- /blog/ai-video-agency-handoff-checklist : AI Video Agency Handoff Checklist (So Clients Don’t Rewrite Your Brief) — A practical AI video agency handoff checklist: start frame approval, motion version log, review gate, deliverables, and usage notes so clients approve faster.
- /blog/ai-image-to-video-guide : Image to Video AI: Turn a Photo Into a Video (2026) — Image to video AI explained: prompted motion vs motion control, when to use each, and how to turn a photo into a video on Zorq AI from 35 credits with free signup credits.
- /blog/5-shot-motion-control-recipe-product-demo : The 5-Shot Motion Control Recipe for a Clean AI Product Demo — Ship a product demo that feels intentional—not random. Use this 5-shot motion-control recipe, a still-first workflow, and a simple review loop to lock consistency fast.
- /blog/how-to-use-kling-motion-control : How to Use Kling Motion Control: Step-by-Step Guide — How to use Kling Motion Control: animate a photo to follow a reference motion clip. Step-by-step setup, key settings, tips, and real credit costs on Zorq AI.
- /blog/kling-ai-pricing-2026 : Kling AI Pricing in 2026: Plans, Credits, Real Costs — Kling AI pricing explained for 2026: free tier, paid plans (~$10–$180/mo), how credits map to videos, and the real per-generation cost of running Kling models on Zorq AI.
- /blog/nano-banana-2-image-generator : Nano Banana 2: What It Is and How to Use It (2026) — Nano Banana 2 is Google's new Gemini image model (3.1 Flash Image): 4K output, accurate text, multi-character consistency. What it does and how to use it from 7 credits on Zorq AI.
- /blog/kling-ai-free-credits-2026 : Kling AI Free Credits in 2026: What's Real, What's Not — Is Kling AI free in 2026? What the free tier really gives you, why the 'unlimited free credits' trick is risky, and how to run Kling models free on Zorq AI.

## Authenticated Pages
- /app/video : Video generation workspace (requires sign-in)
- /app/image : Image generation workspace (requires sign-in)
- /app/history : Generation history for images and videos

## Crawl Guidance
- Prefer canonical URLs.
- Policy pages are public and indexable.
- Blog detail pages live under /blog/{slug}.
- Prioritize public marketing, /video, /image, /pricing, and blog content.
- The public /video and /image pages are for discovery; authenticated workspaces live under /app/*.
- Main product and marketing content is server-rendered HTML.