AI Video Pipeline

🚀 Generate Pipeline

Total Scenes

Completed

Failed

Pending

Voice

Pending

Elapsed

0:00

Active Slots

0 / 6 active

Reference

No reference uploaded

Ready

Load scenes from Batch/JSON tab first, then click Start All

Images done? Click to generate voice + create video-data.json

📋 Batch / JSON

🎬 Scene Data

Paste the scene JSON from Claude/ChatGPT. Format: {"scenes": [{"sceneNumber": 1, "narration": "...", "imagePrompt": "...", "duration": 5}]}

Scene JSON

🎭 Reference Character (Optional)

Click to upload reference character image

Same character in ALL generated scenes

🔮 On Replicate, the reference only applies to FLUX Kontext Pro / Max models (pick one in the Generate dropdown). Other models ignore it.

Character Description — strongly recommended for FLUX Kontext: name skin tone, hair, age & art style

🎙️ Voice Selection

Voice Engine

Voice Model

📁 Project History

Total Projects

Completed

Processing

Failed

⚙️ Settings ✓ Auto-saved

▶ 🎙️ ElevenLabs Voice Not Validated

Get API key from elevenlabs.io/app/settings/api-keys

ElevenLabs API Key

Voice Model

📚 Voice Library — thousands of shared voices; search & Add (then it appears in the dropdown)

➕ Add by Voice ID:

▶ 💎 Azure FLUX 2 Pro (API key + endpoint) Not Saved

API key + endpoint for the 💎 FLUX 2 Pro (Azure AI) image provider — fast text-to-image via Azure AI Foundry's Black Forest Labs deployment. No reference character.

Endpoint URL *

API Key *

▶ 🖼️ Azure FLUX.1 Kontext Pro (reference image) Not Saved

API key + endpoint for 🖼️ FLUX.1 Kontext Pro on Azure AI Foundry — a reference-character model. Upload an image in the Reference Character card, then pick this provider in the Generate dropdown and every scene keeps that same character. Uses the OpenAI-style /openai/v1/images/generations endpoint.

Endpoint URL *

API Key *

▶ 🔮 Replicate API Token (multi-model) Not Saved

API token for the 🔮 Replicate image provider — run almost any model (FLUX, Imagen 4, Ideogram, Recraft, SDXL, FLUX Kontext for reference characters) from one key. Pick the model in the Generate tab dropdown. Get a token from replicate.com/account/api-tokens (pay-as-you-go). Paste it below and save.

Replicate API Token *

▶ 🍌 Gemini Nano Banana (reference character) Not Saved

API key for 🍌 Google Gemini "Nano Banana" image models — the best at keeping the same character across scenes from a reference image. Upload a reference in the Reference Character card, pick this provider + a model in the Generate dropdown, and every scene keeps that character. Get a free key from aistudio.google.com/app/apikey.

Gemini API Key *

▶ 🤖 gpt-image-2 (Azure) — reference image Not Saved

API key + endpoint for 🤖 gpt-image-2 on Azure AI Foundry. Supports text-to-image AND reference editing — upload a reference in the Reference Character card and every scene reuses that character. Uses the OpenAI-style /openai/v1/images/generations + /edits endpoints.

Har API — apna Endpoint + Key * (jitni chaaho add karo — sab pe requests rotate hongi, RPM jud jaayega)

▶ ⚡ Z-Image Turbo (Azure — dedicated GPU) Not Saved

Azure managed-compute image model (Tongyi z-image-turbo etc). Dedicated GPU = koi 30-RPM limit nahi, bahut fast + high concurrency. Endpoint chat wala ho ya images wala — dono chalega (auto-normalize hota hai).

Model / Deployment name

Endpoint URL *

API Key *

▶ 🎨 Qwen-Image (Azure H100 — dedicated GPU) Not Saved

Azure managed-compute image model (qwen--qwen-image-2512) on a dedicated H100 GPU. Koi RPM limit nahi — speed ~14 sec/image (~4-5/min). Concurrency 2-3 rakho (1 GPU hai). Endpoint chat wala ho ya images wala — dono chalega (auto-normalize).

Model / Deployment name

Endpoint URL *

API Key *

▶ 🤖 OpenAI gpt-image (direct) Not Saved

API key for 🤖 OpenAI gpt-image using your own OpenAI account (api.openai.com, direct — no Azure). Supports text-to-image AND reference editing — upload a reference in the Reference Character card and every scene reuses that character. Pick 🤖 OpenAI gpt-image-2 (direct) in the Generate dropdown. Get a key from platform.openai.com/api-keys.

Model

OpenAI API Key *

▶ 🧠 OpenAI Prompt-Fix (Auto-Retry AI) Not Saved

OpenAI key for the auto-fix step: when an image fails to generate even after the retry attempts, this text model analyzes the prompt, detects what caused the content-policy violation, removes/replaces it, and the image is generated again with the cleaned prompt. Used by Scheduled jobs (when "Auto-fix" is on). If left empty, the auto-fix falls back to your image OpenAI key.

Text model (OpenAI: gpt-4o-mini · Azure: your DEPLOYMENT name)

Endpoint URL (only for Azure — leave EMPTY for direct OpenAI sk- keys)

API Key * (OpenAI sk-… or your Azure key)

▶ 📝 Scripts Generator (AI Script Agent) Not Saved

Ye model Topic se poori video script likhta hai (narration + image prompts) — Schedule tab ke 🤖 AI Script Agent me use hota hai. Khaali chhodo to ye tumhari Prompt-Fix key use karega.
Recommended model: gpt-5.4 (tumhare Azure pe deployed).

Model name (Azure DEPLOYMENT name — e.g. gpt-5.4)

Endpoint URL (Azure — direct OpenAI sk- keys ke liye khaali chhodo)

API Key *

▶ 🎬 Render Settings

Resolution

Frame Rate

Format

Transition

Transition Duration: 0.5s

Zoom Effect

▶ 🔧 Advanced

Concurrent Image Generation — higher = faster but hits rate limits sooner. Add more cookie profiles before pushing past ~10.

Image Retry Attempts

👥 Users

▶ Add User

Username

Password

Role

▶ Existing Users

Username	Role	Created

📅 Scheduled Projects

Auto-runs the full pipeline (images → voice → render) at the scheduled IST time

🤖 AI Script Agent — Full Video from a Topic

Type your topics (one line = one video). We write the full script, images & voice and build each project automatically. Turn on Auto-Run for hands-free render → download.

🎭 Reference Image (optional — every scene uses this character)

📎 Click to select reference image

Topics (one line = one video topic — bulk supported)

📝 Script Prompt Template (how the script is written — hook to end; blank = default)

🎨 Image Prompt Template (image style; blank = default)

Image provider

Voice engine

Voice

📚 Voice Library — search & add more voices

🎶 Background Music (optional — plays under the voice)

⬆ Upload

Music volume (15%)

⚡ Turn Auto-Run ON — as soon as the script is ready, run render + download automatically

➕ Add Scheduled Project

Paste scenes JSON, choose provider/model/voice, pick a schedule time (IST) or run immediately. The server runs images → voice → render automatically and makes the video available for download.
Note: browser auto-download is not possible for server-initiated renders; click Download in the job list when done.

Project Name

Schedule Time (IST) — leave blank to use Run Now

Scenes JSON — paste {"scenes":[...]} or an array directly

Image Provider

Model (only for Gemini / Replicate)

— not applicable for this provider —

Voice Engine

Voice

Render Type

Resolution

FPS

Concurrency (image slots)

Retries per scene

Character Description (optional)

Auto-fix failing prompts: after retries fail, OpenAI analyzes the prompt, removes the policy violation & regenerates (set the OpenAI Prompt-Fix key in Settings)

Auto-fix model (text model for prompt cleaning; blank = use Settings)

📋 Scheduled Jobs Auto-refreshes every 10 s

Loading jobs…

💳 Plans & Credits

Manage your subscription & video credits

—

FREE PLAN

Available Credits

50 credits

1 word = 1 credit · ~150 credits = 1 min video
🔵 GPT-Image-2 (premium) = 2× credits

Choose your plan

Zyada credits, kam price — jitna bada plan utna sasta per credit. Cancel anytime.

🚀 Starter

$12/mo

~$0.60/min

3,000 credits

~20 min · 4 videos (5 min)

✅ 1 word = 1 credit
✅ All languages (Hindi, English, Urdu…)
✅ Consistent character (same face)
✅ Voiceover with emotions
✅ No watermark
✅ 5GB storage · Email support

🔥 POPULAR

📦 Standard

$29/mo

~$0.41/min

9,000 credits

~60 min · 12 videos (5 min)

✅ Sab Starter features
✅ 3× credits, better value
✅ Auto Pan & Zoom · Image→Video
✅ Background music & SFX
✅ 20GB storage · Email support

⭐ RECOMMENDED

💎 Pro

$59/mo

~$0.37/min

24,000 credits

~160 min · 32 videos (5 min)

✅ Sab Standard features
✅ Priority generation (2× faster)
✅ Thumbnail generator
✅ 50GB storage
✅ Chat + Email support

👑 BOSS

🏆 Max

$99/mo

~$0.27/min

55,000 credits

~367 min · 73 videos (5 min)

✅ Sab Pro features
✅ Priority generation (5× faster)
✅ Multi-key pooling (fast bulk)
✅ 200GB · Dedicated manager
✅ Early access · Priority support

🔐 Secure payments · 💳 Razorpay / UPI · ⚡ Credits instant add · 🚪 Cancel anytime

🖼️ Thumbnail Generator

gpt-image-2 (Azure) se direct thumbnail generate + download

🖼️ AI Thumbnail — template + script + character se

AI pehle tumhara thumbnail template (style), script (kis baare me), aur reference character detail analyze karega → fir itne thumbnail prompts banayega jitne tum bologe → har ek ko gpt-image-2 (Azure) se direct generate karega → download.

🏷️ Project Title (is naam se history me save hoga)

📝 Thumbnail Prompt Template (style kaisa ho — bold text, face, colors…; khaali = default)

🎬 Script / Topic (video kis baare me hai — AI isse hook samjhega)

🧑 Reference Character Detail (har thumbnail me ye character aayega; khaali = koi character nahi)

🖼️ Image Model (koi bhi provider — busy gpt-image ki jagah dusra chuno)

Kitne thumbnails?

Reference image (optional — same face rakhne ke liye)

Generated Thumbnails

Abhi koi thumbnail nahi — upar se generate karo.

📁 Thumbnail Projects (History)

Abhi koi saved project nahi.

TubeGenix

💳 Plans & Credits

Choose your plan

✏️ Scene — Prompt Edit

AI Video Pipeline

TubeGenix

💳 Plans & Credits

Choose your plan

Edit Scene

✏️ Scene — Prompt Edit