39 AI services & 80+ models in one queue
Kling
video
Gemini
video
Veo
video
Runway
video
Luma
video
MiniMax
video
Seedance
video
Wan
video
HappyHorse
video
Grok Imagine
video
Vidu
video
Pika
video
LTX
video
GPT Image
image
Nano Banana
image
Imagen 4
image
Grok Image
image
Seedream
image
Flux 2
image
fal
gateway
Replicate
gateway
OpenRouter
gateway
Three things that go wrong by shot 46.
Each one costs credits and a morning. Each one has its own screen here.
Check
Drag it. The lock is the difference.
The same shot, the same prompt, the same service. On the left it ran without the reference lock; on the right the lock was in the prompt and the consistency check scored what came back.
Darker hair, an orange shirt: drift, caught before the next shot is paid for.
Eight video projects, eight different briefs.
Explore how a 24-second video can answer a product question, teach a skill, or tell a story. Each example has an English voiceover and a script you can read.
One tab per clip, or one queue per episode.
Your week on a service’s site: a prompt, a wait, a download, a rename — forty-six times, on each site. Here the unit of work is the episode.
Your week in a browser tab
5 separate sites · 46 copy-pastes
+41 tabs
Kling 4.0
Re-type #12
-15 cr
Veo 3.1
Drift 68%!
-20 cr
Runway
Wait 04:18
-18 cr
Luma
Typing...
-14 cr
MiniMax
Quota cap
-12 cr
-79 credits
1 sheet
In Scena Studio
✓ Auto-batched · 4 parallel threads
Video queue — EE012
LIVE
46 shots · 4 threads
SC03 · 1
98% pass
Done ✓
SC04 · 2
97% pass
Done ✓
SC05 · 1
99% pass
Done ✓
SC06 · 2
96% pass
Running 74%
SC07 · 1
Locked
Queued
🔒 Ref Lock: 98.2% avg match
43 / 46 shots done · $4.82
Type the prompt again. Forty-six times, on each site.One sheet, one queue, 39 AI services (80+ models), a model per shot.
Watch the tab. Close it and the batch is gone.Sixty renders overnight; the queue lives on disk.
Describe NOAH again on every site, and watch him drift.NOAH is NOAH on every service, in every episode — and every clip is scored.
Entry plans stop at 10 to 22 clips a month. An episode is sixty.One price per shot at the service’s own rate, shown before you run.
Full comparisonFrom script sheet to finished episode, in one window.
Most AI series are stitched together from four or five tools. Here each job is a screen of the same project.
Queue an episode, walk away.
Paste prompts or import the script sheet, pick a service per shot, and let the queue run. Threads, retries, row delay and per-service rate limits are handled for you; the status bar shows what every provider has spent.
Kling 4.0/O3, Gemini Omni 1.1, Veo 3.1, Runway Gen-4.5/Aleph 2.0, Luma Ray 3.2, MiniMax H3, Seedance 2.5, Wan 3.0, HappyHorse 1.1, Grok Imagine 1.5, Vidu Q4, Pika 2.5, LTX-2.5 — plus 12 image services (47 models)
Text → video, image → video, first-last frame, reference and extend modes
Change the service for selected jobs without retyping a single prompt
Lives in Video queue and Image queue
One cast, every aspect ratio, as many series as you run.
A channel is a hundred videos that must look like one show, on three platforms with three shapes.
YouTube16:9
TikTok, Shorts, Reels9:16
Feed1:1
Every episode is its own projectThe cast carries across seriesThe sheet stays the source of truth
From a fresh install to a finished episode, in eight steps.
Install, sign in once, and everything happens in one window.
STAGE 01
Setup · 60 sec
Install and sign in
Install Scena Studio and configure a supported API service.
scena-studio-1.1.0.exeNo Scena usage quota
Step guide →
STAGE 02
Project Studio
Create the episode
A code, a title, a folder, an aspect ratio.
16:9 YouTube9:16 Shorts1:1 Feed
Step guide →
STAGE 03
Script Sheet
Import the script sheet
Map the columns once; one row becomes one job.
CSV / XLSX Import1 Row = 1 Shot Job
Step guide →
STAGE 04
Reference Sets
Build the reference sets
A code and one image per character, outfit and place.
CHAR LockOUTFIT LockPLACE Lock
Step guide →
STAGE 05
Pre-flight Engine
Set the continuity rules
What may never share a shot. BLOCK stops it on your PC.
BLOCK ($0.00 Spent)WARNAuto-Hold
Step guide →
STAGE 06
Batch Queue
Queue the images, then the video
Stills first, then video; a model per shot.
39 AI Services · 80+ Models4 Parallel Threads
Step guide →
STAGE 07
Local Vulkan GPU
Check every clip before you keep it
Scored against its locks on your GPU; below the mark, re-queued.
Face & Outfit ScoreAuto Re-queue
Step guide →
STAGE 08
Deliver & Lock
Approve, freeze, export
Approve, freeze, export in three shapes.
4K MP4Premiere XMLSRT / WAV
Step guide →
The first run takes about ten minutes, mostly building the reference sets. The next episode starts from the last one.
Download and start
Everything it does, in eight groups.
Choose a group to explore its screens and controls.
Scena StudioPick a group to see its screen.
39 AI services & 80+ models on your own keys. No credits to run out.
Each shot picks its own model at the service’s API price. No credit packs that expire, no markup, and when a better model ships you move the episode, not your account.
Video services
13
Text → video, image → video, first–last frame, reference and extend, depending on what each one supports.
Kling
direct key
Gemini
direct key
Veo
direct key
Runway
direct key
Luma
direct / gateway
MiniMax
direct / gateway
Seedance
gateway
Wan
gateway
HappyHorse
gateway
Grok Imagine
direct key
Vidu
direct key
Pika
gateway
LTX
gateway
Image services
12
Character sheets, backgrounds, props and first frames — 47 models across 12 services in the same queue engine as video.
GPT Image
direct key
Nano Banana
direct key
Imagen 4
direct key
Grok Image
direct key
Seedream
gateway
Flux 2
gateway
Ideogram 3.0
direct / gateway
Recraft V3
direct / gateway
Midjourney V7
gateway
Stability AI (SD 3.5)
direct / gateway
Leonardo.Ai
direct key
Luma Photon
direct / gateway
Gateways & custom
4 + any
One key for many models, or point the app at any OpenAI-compatible endpoint of your own.
fal
gateway
Replicate
gateway
OpenRouter
gateway
WaveSpeed
gateway
Custom provider
your base URL
No markup, ever
You pay each company its own price; the app takes nothing.
Keys stay on your PC
Encrypted with Windows DPAPI. Never uploaded.
Several accounts per service
A flat key fails over instead of stopping the queue.
Not locked to anyone
Re-point jobs to another company; prompt, seed and references stay.
39 services wired in by name — plus any custom REST endpoint
Anything a gateway carries runs today — fal, Replicate, OpenRouter, WaveSpeed, or a custom base URL.
It is not a video editor
No timeline, no object removal. Premiere Pro, DaVinci Resolve or CapCut keeps its job; Scena Studio fills it with footage that already matches.
Open beta · running now
Try the Windows app in the Open Test.
Shared API access is free at the start on enabled services.
During the beta
No usage quota from Scena Studio
Scena Studio covers shared API use at the start of the Open Test on enabled services and sets no per-user quota on that usage. Provider rate limits and availability apply. Once shared access ends, you must use your own API keys. Your provider may charge for usage on your own keys. Render with the providers and models supported in your build Scena Studio requires no card for Open Test Your own keys still work: drop them in and the app uses yours first Provider rate limits and service availability apply to shared keys and your own keys.
Forever
A free Studio lifetime key for testers
Use the tool and send one thing back — a bug, a log, a missing feature — and keep Studio free for the whole 1.x line. One usable bug report is enough Studio, on three PCs, through every 1.x version Testers are credited in the release history, if they want to be No invoice, no renewal, no seat count.
Download and run it
Set up your API key once for each provider you use.
Put it through a real episode
A real scene, with locks, retries and a service that fails.
Send one report
Help → Report a problem, the report page, or the group. Anything usable earns the key.
Download the beta
Join the testers’ group
Provider rate limits and availability apply. Shared keys are available only where enabled.
What early episode creators report from the queue.
Verified notes from open-beta testers running multi-shot animated episodes on Windows — with exact model routing, API cost, and drift logs.
✓ VERIFIED LOG
Marcus K. · Kids Animation Channel · v1.1.0
Kept our 5-character family cast on-model across 48 shots on Kling 4.0 and Veo 3.1. Local GPU drift check caught 3 off-model frames automatically — 98.2% face match and full YouTube monetization approval.
✓ VERIFIED LOG
Elena R. · Indie Series Director · v1.1.0
Cut our 8-minute episode render bill from $95/mo in expiring web credits down to $6.40 on direct fal.ai and Google keys. The pre-queue BLOCK rule stopped 4 unreferenced shots before spending a cent, with zero watermarks.
✓ VERIFIED LOG
David T. · Faceless Storytelling Studio · v1.1.0
Imported a 60-row CSV script sheet, locked character turnarounds once, and let the Windows queue run overnight. Woke up to 60 finished 4K clips, lip-sync audio, and a Premiere XML timeline — no browser tabs crashing.
✓ VERIFIED LOG
Hai D. · Educational Cartoon Producer · v1.1.0
Dubbed our 46-shot episode into English, Vietnamese, Spanish, Japanese, and Korean in one afternoon. Locked ElevenLabs voice IDs per character and ran Real-ESRGAN 4K upscaling 100% free on our RTX 4070.
Every report above is pulled from verified v1.1.0 project logs — testers who submit a complete episode log or reproducible bug report keep a free lifetime Studio key.
19
video and image services in one queue
15
interface languages, switched without a restart
$0
cloud cost — upscaling, checks and storage run on your PC
40
screens — no second app, no spreadsheet
What leaves your PC, and what never does.
Press RUN and three things go to three places. This is the whole list.
Stays on your PC
Your episodes: prompts, references, renders, the sheet
Your API keys, encrypted with Windows DPAPI
The queue and its history, on disk
Consistency scores, from your GPU
Goes to the AI service you chose, for that one shot
That shot’s prompt
The references that shot uses
The pooled key, or your own for that service
Nothing from any other shot or episode
Goes to your Scena Studio account
Your email and display name
The licence and which PCs use it
Four monthly counters per service
Never a prompt, an image or a render
We cannot see your episodes. We can see that you rendered 212 minutes in October.
Before you download.
If something here is still unclear, the group answers faster than email.
Anything else — a service that is not in the list, a licence question, a bug — goes to the group or the in-app report form.
Community & support
Scena Studio covers shared API use at the start of the Open Test on enabled services and sets no per-user quota on that usage. Provider rate limits and availability apply. Once shared access ends, you must use your own API keys. Your provider may charge for usage on your own keys.