Ovi
Open-source, uncensored image-to-video model with synchronized audio generation — runs privately on Venice with zero retention.
Overview
What is Ovi
Ovi is an open-source image-to-video model developed by Character.AI that generates 5-second video clips with synchronized audio from a single image and text prompt. Released in September 2025, it uses twin-DiT backbones for cross-modal audio-video fusion and supports private generation on Venice with no prompt retention.
Running it privately on Venice
On Venice, Ovi runs with full privacy — your prompts and inputs are never stored or used for training. As an open-source, uncensored model, it enables permissionless video creation with audio, ideal for creators who value sovereignty and transparency. You pay only per clip, with no subscription required.
Assessment
Strengths and limitations
- Generates synchronized video and audio in a single pass using blockwise cross-modal fusion.
- Open-source and uncensored: fully self-hostable with Apache-2.0 licensing.
- High-quality audio generation with realistic speech and sound effects via dedicated audio tower.
- Supports fine-grained control via text prompts including audio descriptions.
- Runs privately on Venice with zero data retention — ideal for sensitive or commercial use.
- Limited to 5-second clips (10 seconds in Ovi 1.1) — shorter than some competitors.
- No native support for resolutions above 960×960 without upscaling.
- Less polished cinematic output compared to closed models like Veo 4 or Sora.
- Requires technical setup for local deployment; less accessible than web-only tools.
Samples
Sample outputs
Generated on Venice with our standard prompt suite — the same prompts we run through every model of this type, so you can judge it like-for-like.
Slow aerial drone shot gliding over a misty mountain valley at golden hour, sunlight piercing clouds onto a winding river, ancient pine forests on either side, ultra-smooth motion, professional color grading, atmospheric haze, 4K
Calm ocean waves rolling onto a black-sand beach at sunrise, golden light on wet sand, foam dissolving into the shore, a single silhouetted figure at the waterline, smooth continuous forward push, serene cinematic atmosphere
A slow tracking shot through a rain-soaked Tokyo street at night, neon reflecting in puddles, steam rising from a food cart, a person with a translucent umbrella, shallow depth of field, teal-and-orange grade, smooth steady camera
Capabilities
What it supports
- Text to video
- Image to video
- Reference to video
- Native audio generation
Specifications
Datasheet
- Maker
- Character.AI
- Released
- September 2025
- Architecture
- Twin-backbone Diffusion Transformer (Wan2.2-based, A14B MoE)
- Modality
- Image-to-video with synchronized audio
- Clip length
- 5s (10s in Ovi 1.1)
- Clip lengths
- 5s
- Mode
- image-to-video
- Audio
- Yes
- Privacy on Venice
- Private — zero retention
- Available on Venice since
- Sep 2025
- License
- Apache-2.0
API
Call it from your code
Venice exposes this model through the REST API. Queue a generation with the model id.
curl https://api.venice.ai/api/v1/video/queue \
-H "Authorization: Bearer $VENICE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "ovi-image-to-video",
"prompt": "Aerial drone shot over a misty mountain valley at golden hour"
}'
# Use the returned queue_id with https://api.venice.ai/api/v1/video/retrieve.
# Call /video/complete after downloading if needed.Pricing
What it costs on Venice
Pay per clip on Venice — price scales with resolution and duration (5s–5s), from $0.22.
New Venice accounts include a free daily allowance and 500 welcome credits — no credit card required.
Alternatives
How it compares
| Model | Best for | Max clip length | Audio support | Open weights | Price (Venice) |
|---|---|---|---|---|---|
| Ovi | Open-source, audio-synchronized image-to-video model with private generation on Venice. | 5s (10s in 1.1) | Yes | Yes | from $0.22 |
| Wan 2.7 | Open-source video model, but lacks native audio generation. | — | — | Yes | from $0.55 |
| Wan 2.7 Enhanced | Enhanced version of Wan with better visual fidelity, still no audio. | — | — | Yes | from $0.68 |
| Kling O3 Pro | High-end proprietary model with long-form video, but no open access. | — | — | No | from $0.46 |
Open-source, audio-synchronized image-to-video model with private generation on Venice.
Use cases
What it is good for
- 01Creating short-form social media videos with custom audio and visuals.
- 02Prototyping animated content for marketing or storytelling.
- 03Generating sound-rich video clips for games or interactive media.
- 04Privacy-sensitive video generation where data sovereignty is critical.
- 05Research and development in multimodal audio-video synthesis.
Prompting
Getting better results
Use the format 'Audio: [description]' to specify sound design in your prompt.
Start with clear, simple image inputs — detailed stills yield better motion.
Describe both visual action and audio cues to maximize synchronization.
For longer clips, use Ovi 1.1 which supports 10-second generation at 960×960.
Version history
Initial release — 5-second image-to-video with audio.
Updated to support 10-second clips at 960×960 resolution.
FAQ
Frequently asked questions
Ovi is an open-source image-to-video model by Character.AI that generates 5-second video clips with synchronized audio from a single image and text prompt. It uses twin-DiT backbones for cross-modal fusion and supports private, zero-retention generation on Venice.
On Venice, Ovi costs from $0.22 per 5-second clip, with pricing scaling based on resolution and duration. You pay per clip — no subscription is required.
Yes. Ovi is fully open-source under the Apache-2.0 license and can be self-hosted for free. On Venice, you pay only for usage, with no mandatory subscription or hidden fees.
Yes. Ovi simultaneously generates synchronized audio and video. It includes a dedicated 5B-parameter audio tower trained from scratch to produce realistic speech and sound effects.
Yes. On Venice, Ovi runs with zero retention — your prompts and inputs are not stored or used for training. This makes it ideal for private, secure, or commercial video creation.
Sora produces longer, more cinematic videos but is closed and not privacy-guaranteed. Ovi is open-source, uncensored, and runs privately on Venice, making it better for sovereignty and customization, though limited to shorter clips.
Ovi natively supports 960×960 resolution. The Ovi 1.1 update enables 10-second clips at this resolution, trained specifically on high-resolution data.
No. Ovi is a generative model focused on image-to-video and audio synthesis. It does not support external tool use or real-time web search during generation.
Run Ovi privately
No prompt logging. No data used for training.