…without the access code. 🧙 BOP Studio — photo to video, private access.
BOP Studio
KIE …WS …RH …Kling …
Agent
Factory
Connecting to the studio…
Loading…
No recent sync from the PC. Check its internet connection. The Factory may still be running locally; status updates resume automatically.
To reviewReview studio
Nothing to review right now. New movement pilots appear here automatically.
0:00 / 0:00
Loops
Changes apply at the next PC sync (within 5 min).
Stopped loops (0)
No loop yet. A loop downloads Instagram accounts and turns their reels into videos for your creators, on its own.
▸ Add videos Instagram links, usernames or Google Drive
To keep downloading accounts on their own, start a loop.
▸ Version outfits the clothes versions pick from
Versions pick different outfits from this list. Take 1 always keeps the original outfit.
⬆ Upscale a video
Super-resolution only — nothing is re-generated, nothing is re-animated. The video
keeps its exact content and comes back sharper.
1 Which video?
No finished video yet — make one in Create videos, or upload a file.
🎞️Drop videos here or click to browse MP4 / MOV / WebM — 50 MB per file
It has to be a direct file link the upscaler can download — not a player page.
2 How sharp?
3 Upscaler
3 Where does it go?
Saved to Google Drive → BOP STUDIO \ today's date
🤖 Boptranscribe
Paste Instagram links (one per line) — your PC downloads each video, transcribes the
audio and writes the full recreation script + the on-screen text, for free.
Results land in 📼 IG scripts.
Your PC grabs the first frame of the reel, the studio puts her in it with Seedream 5.0 Pro (see Create photos), and the video starts by itself once the script is written. Nothing ticked = plain analysis as usual.
Nothing in the queue — paste a link above. The finished scripts are in 📼 IG scripts.
⬇ Download an account's videos
Type Instagram usernames: your PC downloads every video of each account into
its archive folder on the Drive. Open accounts are fetched for free;
restricted ones (18+, login-only) go through HikerAPI, about one API call per 30 videos.
Already-downloaded videos are skipped, so you can run the same account again later to get the new ones.
Nothing yet — type a username above. Files land in Stolen videos \ username on your Drive.
📼 IG scripts
Videos analyzed on your PC by Boptranscribe : paste an Instagram link there,
the recreation script lands here with its thumbnail — copy it, or send it straight
to the prompt box of Create videos.
Nothing yet — paste an Instagram link in Boptranscribe on your PC. 🤖
💬 Add dynamic subtitles
Instagram-style captions: every spoken word pops on screen, the current word gets its
own color or box. The speech is detected on your PC (free), you fix the text,
pick a style, and a new video is saved — the original is never touched.
⚠ The desktop agent is offline — word detection runs on your PC, start the agent first.
1 Which video?
No finished video yet — make one in Create videos, or upload a file.
🎞️Drop videos here or click to browse
⚡ straight to your PC MP4 / MOV / WebM — 50 MB per file
📁 Big files, free: your Drive inboxno credits, any size
Copy the video into G:\My Drive\BOP STUDIO\subtitles-inbox — from this PC,
the Google Drive app on your phone, or drive.google.com. It shows up here about a minute after the Drive
sync; tick it and go. The subtitled video lands in the usual Drive folder and the original moves to
subtitles-inbox \ done.
Nothing in the inbox yet.
It has to be a direct file link your PC can download — not a player page.
2 Speech detection
Runs on your PC with Whisper — free, no credits spent. You will review and fix the text before anything is burned in.
3 Where does the subtitled video go?
Saved to Google Drive → BOP STUDIO \ today's date
Free — everything runs on your PC.
1 Who is this for?
2Photo(s) to edit
🖼️Drop photos here or click to browse JPG / PNG / WebP — one result per photo, up to 12
or
The photo you drop is the one that gets changed. Say what to change in the prompt step — the model keeps the rest.
Her ticked library photos travel with each photo you edit — tick them in ✨ New photo of her.
OFF: 3 photos → 3 results. ON: all photos go into ONE result — e.g. the dress from photo 2 on the girl from photo 1.
2 Her reference photos
These photos tell the model what she looks like — they are never shown in the result.
Your description says what the new photo shows. Best: 2 to 4 photos, one face close-up
+ one full body, good light, no sunglasses.
🖼️Drop photos of her here or click to browse Used for this photo only — not added to her library
2 Which photo?
No finished photo yet — make one first, or upload a file.
🖼️Drop photos here or click to browse JPG / PNG / WebP — sent at full size when under 4 MB, up to 12
Nothing is re-drawn: the photo keeps its exact content. Sharpen makes it bigger and crisper; Remove background cuts her out on a transparent PNG.
3 What to do
3 Pick a style
🔒 This model keeps its own safety filter: it refuses nudity, lingerie close-ups and explicit
words — and still charges for the try. For spicy scenes pick a 🔓 model.
KIE
Rule of thumb: bold scene → Seedream 5.0 Pro · her exact face in a safe scene → Nano Banana Pro · many quick tests → Z-Image or Seedream 5.0 Lite.
Vertical for Stories and Reels, square for the feed. Standard is enough for chat previews, High for posts, Ultra only if you'll crop or print.
Options
4Describe the photo
One click writes a ready prompt below — edit it or hit Create.
Enter adds a line · Ctrl+Enter writes the prompt
Your text. The studio adds one line about her face (from the reference photos) and the camera look automatically — the full prompt shows on the card.
The camera look is added at the end of your prompt so every photo keeps the same feel. “Photos to make” runs the same prompt several times — each result comes out different.
5 Sharpen before saving optional
Runs on the finished photo. Turn it on when you generated at 1K to save money and want it big anyway.
⚠ The desktop agent is offline — photos are still made and kept here; they go to Google Drive when it comes back.
Your settings are remembered on this PC.
1 Who is this for?
Saved to Google Drive → BOP STUDIO \ today's date
2 Reference media
These photos say who she is — they are never animated as-is. Your prompt says
what happens, and one video is built from it.
2Add your photos
🖼️Drop photos here or click to browse JPG / PNG / WebP — one video per photo
3 Pick a style
KIE
5s
Advanced options
4 Text prompt
Enter adds a line · Ctrl+Enter writes the prompt
This is what the video model receives.
5 Sharpen before saving optional
3 Who copies the movement?
MCP
4 Movement to copy
No saved motion yet — save one in the Kling app (Motion Control → save the motion).
The desktop agent refreshes this list every 10 minutes.
Motions saved on your Kling account — nothing to upload. Kling uses them directly; WaveSpeed models read the video from Kling's servers.
No movement video yet — Upload one here, or drop it in
G:\My Drive\BOP STUDIO\driving-videos\ (the desktop agent adds it).
RH = on RunningHub (Mimic Motion) · ☁ = copy in the studio (Kling & WaveSpeed). A video dropped in the Drive folder is RunningHub-only: upload it here to use it everywhere.
🎞️Drop movement videos here or click to browse MP4 / MOV / WebM — up to 50 MB each, they join the library for every model. Then « 🖼 First frame » puts her in the video's first image.
A direct file link — not a player page. Kling and WaveSpeed only (RunningHub needs the file).
5 Options
RunningHub settings
The rest waits in the studio queue and starts automatically when a slot frees up.
6 Add subtitles optional
Free — Whisper runs on your PC. The words are detected automatically, then the job
waits for your review: you fix the text, tweak the style and hit Burn.
7 Review before posting recommended
Available with Kling Motion Control and Mimic Motion
Videos wait for your review in My videos. After approval, dialogue is subtitled automatically and finals go to the creator's Drive folders: ANDROID 80% · IPHONE 20%. Videos without a creator use the No creator folder. Each delivered video shows its actual Android or iPhone destination. Raw previews stay in <Creator>\_review\; approved videos go directly into <Creator>\ANDROID\ or <Creator>\IPHONE\. Move published videos into that folder's Posted subfolder.
Claude (Haiku, on your PC) looks at each photo, then the writer above scripts it. Keep this tab open until every video has started.
⚠ This model runs through your Kling subscription on the desktop agent — start the agent first (it is offline right now).
▸ Factory videos The last 7 days, with their status
Your PC runs an older factoryd: the gallery appears after its update
Filters
No video made on this day yet
My photos
Today $0.000 running
0 selected
📸
Your photos will show up here. Pick a creator, click a look, hit Create.
🤖 Automation
Pick a girl and a model: the agent invents a fresh scenario, writes the prompt and
submits a video — over and over, until the count is reached. Her saved reference
photos are used every time. Each creator runs her own series, in parallel — three videos in flight per series.
ON
Each girl runs her own series — they go in parallel.
📊 Analytics
My videos
0 running
🎬
Your videos will show up here. Add photos and hit Generate.
🪪 Reference photos
Save each girl's photos once — they are reused in every “Reference → scene” video, so
you never drop the same files twice. Best pair: one vertical full-body shot and one
vertical face close-up, neutral expression, face filling about ⅔ of the frame.
Flow uses the first-frame screenshot and its scene description together with your creator’s existing references. The result becomes the photo for Copy a movement.
Extra frames keep everything (pose, place, light) and only recolor her outfit — one photo job each, and one Copy a movement each when you hit Generate.
Import every available video. New movements first become a Vicky pilot in Tinder; choose models and video counts after approval. Existing videos are skipped.
Advanced settings
Ctrl+Enter start · Esc back
Stop this loop?
Edit loop
Ctrl+Enter save · Esc back
Factory settings
One per line. Each first frame gets one at random (a garment such as tank top or yoga pants); the versions of a video get different ones. Never an explicit word: Google Flow refuses them.
Preparing the preview on the PC (next sync, up to 5 min)…
Delete this video?
Request stopping any active rendering, then delete this video from Factory and Google Drive, including pending copies and retries. Other versions are kept.
The request reaches the PC at its next sync; an offline PC handles it when it returns. The video stays hidden until cleanup is confirmed. This cannot be undone in Factory.