Furkan Gözükara
patreon
Ovi is Local Version of VEO 3 & SORA 2 - The first-ever public, open-source model that generates both VIDEO and synchronized AUDIO, and you can run it on your own computer on Windows even with a 6GB GPUs - Full Tutorial for Windows, RunPod and Massed Compute - Gradio App
🕑 Added 2025-10-11 14:10:53 +0000 UTCTutorial Link on YouTube
Link : https://youtu.be/T00VmkMQRPQ
Ovi is Local Version of VEO 3 & SORA 2 - Even 6 GB GPUs Runs on Windows - Generate Videos With Sound
Info
Forget waiting lists and expensive APIs. The era of closed-off, corporate-controlled AI video generation is soon over. This is Ovi : The first-ever public, open-source model that generates both VIDEO and synchronized AUDIO, and you can run it on your own computer—even with a 6GB GPU! This isn't just a demo; it's a full, step-by-step revolution.
Ovi Pro Premium Download Link
https://www.patreon.com/posts/download-ovi-pro-premium-140393220
Windows Requirements Tutorial
Auxiliary Links
SECourses Discord
SECourses Reddit
My LinkedIn
Tutorial Info
In this ultimate A-Z guide, I'll show you EVERYTHING you need to know to install and master this Sora 2 and VEO3 like AI. We'll go from zero to generating incredible talking videos from text or a single image.
🔥 In This Tutorial, You Will Learn To:
🎓 Master the Ultimate SORA 2 and VEO 3 Alternative: The first true open-source challenger to OpenAI & Google.
💻 Run on Low-Spec Hardware: We've optimized this to run on GPUs with as little as 6GB of VRAM!
💸 Generate for FREE: No credits, no subscriptions. Run it locally on Windows or cheaply in the cloud.
🗣️ Create Synced Audio & Video: Go beyond silent movies. Make your characters speak with perfect lip-sync.
☁️ Install ANYWHERE: Complete one-click install guides for Windows, MassCompute, and RunPod.
🖼️ Animate Any Image: Bring your static images to life with stunning animation and speech.
🚀 Unlock Pro Features: Dive deep into batch processing, video extensions, LoRA support, and advanced optimizations.
🕒 VIDEO CHAPTERS:
0:00 Introduction to OVI: The First Open-Source Audio+Video AI
0:37 Impressive AI Video Generation Demos
1:00 Core Capabilities: Text-to-Video & Image-to-Video Animation
1:26 UI Walkthrough: Uploading Images & Videos
1:39 Auto Cropping, Padding & Aspect Ratio Control
1:53 Adjusting Base & Output Video Resolution
2:23 Using Built-in Examples & Understanding Prompt Structure
2:36 Essential Prompting Syntax: Speaking & Audio Tags
2:49 Built-in Prompt Validation & Syntax Error Checker
3:05 Advanced Feature: Seamless Video Extension & Storytelling
3:52 How Video Extension Uses the Last Frame for Continuity
4:19 Setting Custom Video Duration & FPS Explained
4:38 Using a Video as an Initial Input Frame
4:53 Seed, Disabling Audio & Full Metadata Explained
5:22 How to Use LoRAs with OVI (Video & Sound Layers)
6:38 DEEP DIVE: GPU & Memory Optimization Settings
6:51 Block Swap: Running on Low VRAM GPUs (6GB+)
7:11 CPU Offloading & "Clear All Memory" for Low RAM Systems
7:44 Intelligent Scaled FP8 for VRAM Reduction & Quality
8:25 Tiled VAE Decode: The Key to Low VRAM Performance
8:48 Using the Full Preset System for Different Setups
9:09 Pro Feature: Automated Batch Processing from a Folder
10:32 OVI Installation Guide Introduction (Windows, MassCompute, RunPod)
10:50 Step 1: Download & Extract the Files on Windows
11:12 Step 2: Running the One-Click Installer & Update Script
11:39 CRITICAL: Windows Prerequisite Installation Guide
12:53 Step 3: Using the Resumable Model Downloader
14:52 How to Update the Application
15:06 First Launch & Verification Test
18:18 Pro Tip: Running the App on a Second GPU
20:32 Advanced Prompting Guide: How to Write Effective Prompts
20:53 Using Google Gemini to Generate OVI Prompts (Detailed Walkthrough)
22:02 Pro Tip: Setting Custom Durations Per Prompt Line
23:21 Cloud Guide: How to Install on MassCompute
23:44 Deploying the Machine & Selecting the Right GPU
25:03 Connecting via ThinLinc & Transferring Files
25:45 Running the MassCompute Install Script
28:04 Accessing the App & Performance on MassCompute
29:53 Cloud Guide: How to Install on RunPod
30:21 Configuring the RunPod Pod (Template, Disk, GPU)
31:56 Connecting to JupyterLab & Uploading Files
32:26 Running the RunPod Install & Download Scripts
34:02 Accessing the App on RunPod (Gradio vs Proxy)
38:41 Pro Feature: Using the Gradio Queue System for Batch Jobs
40:45 Final Words, Support & Community Links (Discord, Reddit)
Comments
Furkan Gözükara
i think for this task you can use reguler wan 2.2 : https://youtu.be/c3gEoAyL2IE https://youtu.be/3BFDcO2Ysu4
Volodymyra Malchevska
Hi, thank you for your hard work! How can I generate video with human but without talking (like b-roll). Thank you in advance!
Furkan Gözükara
yes error reason is easy. please set 100 gb virtual disk, restart run and let me know : https://www.windowscentral.com/software-apps/windows-11/how-to-manage-virtual-memory-on-windows-11
James
Hello my friend. First of all, thank you for providing such great information! I ran into an error upon running my first example video. Can you look at the text and determine what I may have missed. I;m new to this, so I appreciate any help. [STARTUP] Set MKLOMP threads to 12 for optimal CPU performance [DEBUG] Parsed args single_generation=False, single_generation_file=False, test=False, test_subprocess=False Starting Gradio interface with lazy loading... Share mode DISABLED (local only) Use --share flag to enable public access with a shareable URL Ovi Pro SECourses Premium App v8.3 [DEBUG] Main block single_generation_file=None, single_generation=None, encode_t5_only=False, test=False, test_subprocess=False Running on local URL http127.0.0.17860 To create a public link, set `share=True` in `launch()`. Auto-loading last used preset 24-GB GPUs - Faster FP8 Scaled ============================================================ GPU DETECTION RESULTS GPU Model NVIDIA GeForce RTX 4090 GPU Count 1 VRAM Size 23.99 GB ============================================================ ============================================================ SYSTEM RAM DETECTION RESULTS Total RAM 15.94 GB ============================================================ ✓ RAM optimization Enabled Clear All Memory (RAM 15.9GB 128GB) [PRESET] Applied automatic optimizations for '24-GB GPUs - Faster FP8 Scaled' RAM 15.9GB 128GB → Enabled Clear All Memory ============================================================ GPU DETECTION RESULTS GPU Model NVIDIA GeForce RTX 4090 GPU Count 1 VRAM Size 23.99 GB ============================================================ ============================================================ SYSTEM RAM DETECTION RESULTS Total RAM 15.94 GB ============================================================ ✓ RAM optimization Enabled Clear All Memory (RAM 15.9GB 128GB) [PRESET] Applied automatic optimizations for '24-GB GPUs - Faster FP8 Scaled' RAM 15.9GB 128GB → Enabled Clear All Memory ============================================================ GPU DETECTION RESULTS GPU Model NVIDIA GeForce RTX 4090 GPU Count 1 VRAM Size 23.99 GB ============================================================ ============================================================ SYSTEM RAM DETECTION RESULTS Total RAM 15.94 GB ============================================================ ✓ RAM optimization Enabled Clear All Memory (RAM 15.9GB 128GB) [PRESET] Applied automatic optimizations for '24-GB GPUs - Faster FP8 Scaled' RAM 15.9GB 128GB → Enabled Clear All Memory [LORA DEBUG] Received from UI lora_1='None', lora_1_scale=1, lora_1_layers='Video Layers' lora_2='None', lora_2_scale=1, lora_2_layers='Video Layers' lora_3='None', lora_3_scale=1, lora_3_layers='Video Layers' lora_4='None', lora_4_scale=1, lora_4_layers='Video Layers' lora_specs=None [LORA STATUS] UI selections will be processed 4 LoRA(s) [1] None (scale 1) [2] None (scale 1) [3] None (scale 1) [4] None (scale 1) [RESOLUTION] Using exact user-specified resolution 960x544 ================================================================================ VIDEO GENERATION STARTED enable_multiline_prompts False enable_video_extension False Text prompt A man in a workshop, with soldering irons and circ... Image path None Resolution 544x960 Base Resolution 720x720 Duration 5 seconds Seed 99 Num generations per prompt 1 Video extensions 0 Valid prompt lines detected 1 ================================================================================ ================================================================================ CLEAR ALL MEMORY ENABLED Main process will NOT load any models All generations will run in separate subprocesses VRAMRAM will be completely cleared between generations ================================================================================ [PROMPT 11] Processing A man in a workshop, with soldering irons and circ... [GENERATION 11] Starting with seed 99 [GENERATION LORA] No LoRAs applied (using base model only) [SUBPROCESS DEBUG] Passing lora_specs to subprocess [] [SUBPROCESS DEBUG] lora_specs type class 'list', length 0 [SUBPROCESS] Running generation in subprocess... [SUBPROCESS] Command COvi_Pro_v8Ovi_ProvenvScriptspython.exe COvi_Pro_v8Ovi_Propremium.py --single-generation-file COvi_Pro_v8Ovi_Protmpr4tva3ao.json [SUBPROCESS] Params file COvi_Pro_v8Ovi_Protmpr4tva3ao.json [DEBUG] Parsed args single_generation=False, single_generation_file=True, test=False, test_subprocess=False Starting Gradio interface with lazy loading... Share mode DISABLED (local only) Use --share flag to enable public access with a shareable URL Ovi Pro SECourses Premium App v8.3 [DEBUG] Main block single_generation_file=COvi_Pro_v8Ovi_Protmpr4tva3ao.json, single_generation=None, encode_t5_only=False, test=False, test_subprocess=False [DEBUG] Taking single_generation_file path [SINGLE-GEN] Loaded params from file COvi_Pro_v8Ovi_Protmpr4tva3ao.json [SINGLE-GEN] Starting generation with params ['text_prompt', 'image', 'video_frame_height', 'video_frame_width', 'video_seed', 'solver_name', 'sample_steps', 'shift', 'video_guidance_scale', 'audio_guidance_scale', 'slg_layer', 'blocks_to_swap', 'video_negative_prompt', 'audio_negative_prompt', 'use_image_gen', 'cpu_offload', 'delete_text_encoder', 'fp8_t5', 'cpu_only_t5', 'fp8_base_model', 'use_sage_attention', 'no_audio', 'no_block_prep', 'num_generations', 'randomize_seed', 'save_metadata', 'aspect_ratio', 'clear_all', 'vae_tiled_decode', 'vae_tile_size', 'vae_tile_overlap', 'base_resolution_width', 'base_resolution_height', 'duration_seconds', 'auto_crop_image', 'base_filename', 'output_dir', 'text_embeddings_cache', 'enable_multiline_prompts', 'enable_video_extension', 'disable_auto_prompt_validation', 'force_exact_resolution', 'lora_specs', 'merge_loras_on_gpu'] [SINGLE-GEN] Text prompt A man in a workshop, with soldering irons and circ... [LORA DEBUG] Received from UI lora_1=None, lora_1_scale=1.0, lora_1_layers='Video Layers' lora_2=None, lora_2_scale=1.0, lora_2_layers='Video Layers' lora_3=None, lora_3_scale=1.0, lora_3_layers='Video Layers' lora_4=None, lora_4_scale=1.0, lora_4_layers='Video Layers' lora_specs=[] [LORA STATUS] No LoRAs selected [LORA] Using pre-built lora_specs from subprocess 0 LoRA(s) [RESOLUTION] Using exact user-specified resolution 960x544 ================================================================================ VIDEO GENERATION STARTED enable_multiline_prompts False enable_video_extension False Text prompt A man in a workshop, with soldering irons and circ... Image path None Resolution 544x960 Base Resolution 720x720 Duration 5 seconds Seed 99 Num generations per prompt 1 Video extensions 0 Valid prompt lines detected 1 ================================================================================ ================================================================================ INITIALIZING OVI FUSION ENGINE IN MAIN PROCESS Block Swap 0 blocks (0 = disabled) CPU Offload True Image Generation False No Block Prep False Note Models will be loaded in main process (Clear All Memory disabled) ================================================================================ Removing weight norm... [OK] OviFusionEngine initialized successfully (models will load on first generation) [PROMPT 11] Processing A man in a workshop, with soldering irons and circ... [GENERATION 11] Starting with seed 99 [GENERATION LORA] No LoRAs applied (using base model only) [DEBUG] No text_embeddings_cache available - T5 will be loaded in-process [SAGE ATTENTION] Enabled - using Sage Attention for ~10% speedup & lower VRAM [T5 CACHE] Cache key 820376c1545da46ef0063d280d415995337c1ff485c20951ef2635895a4dde38 [DEBUG] No cached embeddings - will load T5 ================================================================================ STEP 12 Loading T5 text encoder FIRST to minimize RAM usage ================================================================================ ================================================================================ Loading OVI models for first generation... Block Swap 0 blocks CPU Offload True ================================================================================ Initial VRAM 0.00 GB ================================================================================ SCALED FP8 T5 Loading T5 in Scaled FP8 format Expected VRAM savings ~50% (~5-6GB saved) ================================================================================ [FP8 CACHE] Found cached FP8 checkpoint COvi_Pro_v8Ovi_ProckptsWan2.2-TI2V-5Bmodels_t5_umt5-xxl-enc-fp8_scaled.safetensors [FP8 CACHE] Creating structure on CPU first (avoids BF16 VRAM allocation) [T5 LOAD][FP8] Structure created on CPU in 169.04s (FP8 cached path) WARNINGrootFailed to load cached FP8 checkpoint (The paging file is too small for this operation to complete. (os error 1455)). Rebuilding FP8 weights... [FP8 CACHE] Failed to load cached FP8 checkpoint (The paging file is too small for this operation to complete. (os error 1455)). Regenerating... [FP8 CACHE] No valid cache found - will quantize from scratch [SUBPROCESS] Generation failed with return code 3221225477 [GENERATION 11] No output file found in COvi_Pro_v8Ovi_Prooutputs after retries Total generation time 202.36 seconds ================================================================================ VIDEO GENERATION COMPLETED Final output path None File exists No ================================================================================ [MEMORY CLEANUP] Final cleanup completed - all generation memory freed
Furkan Gözükara
sure let me know. you can email me cmd logs : monstermmorpg@gmail.com
James Charleston II
hello my friend. The latest 8 version seems to be doing better but running a test with the first T2V with Extention did not work properly. It completed the first video, did create a second video, the first video was not created and it fails to merge because the 3rd video is missing, but I notice it dropped out of the venv on my F drive which is why there app is, to the system python on the C drive. I will try to do another run.
Furkan Gözükara
it is so easy. follow this tutorial and install python 3.10.11. my installer will auto generate venv : https://youtu.be/DrhUHnYfwC0
BecauseReasons
Please start using Python 3.12, I don't understand these OLD Python versions; either that or maybe start using Conda to manage installations since you can easily control Python versions that way too.
