MiniMax H3: Free AI Video Generator With Sound | MaxH3

# MaxH3 > Run MiniMax H3, also called Hailuo 3: generate 4-15s clips at native 2K with stereo audio, then change one thing by describing it. Multimodal video editing online. MiniMax H3 AI Video Generator: Every top AI video model, one workspace. ## Video generation - [Text to Video](https://maxh3.io/text-to-video): One prompt box, every major video model behind it. MaxH3's text to video generator holds characters and style steady from shot to shot. Free credits to start. - [Image to Video](https://maxh3.io/image-to-video): Feed MaxH3 a still and it moves — camera pans, subject motion, ten-plus framing options. Your image to video AI runs on every major model. Free credits. - [Reference to Video](https://maxh3.io/reference-to-video): Lock your character once and MaxH3 holds it — same face, same form, every frame. Reference to video that keeps subjects consistent across shots. Free credits. ## Audio generation - [Text to Speech](https://maxh3.io/text-to-speech): Paste a script, pick a voice, get an MP3. MaxH3's text to speech reads your words with natural pacing and pause control. Free credits to start. ## Image generation - [AI Image Generator](https://maxh3.io/ai-image-generator): Free AI image generator that turns text prompts into high-resolution images with GPT Image 2.5, Nano Banana, Seedream and 20+ models. Free credits, no card to start. - [Image to Image](https://maxh3.io/image-to-image-ai): Turn photos into clean, on-brand visuals. MaxH3's image to image AI changes style, lighting and backgrounds with GPT Image 2.5, Seedream and more. - [Photo Editor](https://maxh3.io/ai-photo-editor): Edit photos online with a text prompt. MaxH3's free AI photo editor removes objects, swaps backgrounds, retouches and restyles images — no masks or layers. ## Video models - [Seedance 2.5 AI Video Generator](https://maxh3.io/vm/seedance-2-5): Run ByteDance Seedance 2.5 in the browser. One-take 30-second video with synced audio, up to 50 reference files, and edi