SuperAPI AI Directory
Partner with us中文AI newsFavorites

AI video

2,208 results · Page 29 of 62

MixHub AI

AI video

MixHub AI is a web-based one-stop AI creation platform that aggregates multiple AI models such as GPT-5, Flux, Claude, Qwen Image, Kling, Hailuo, and others, focusing on multi-model AI video generation and image creation. The platform integrates video generation models including Runway, Hailuo (MiniMax), Kling, Luma Dream Machine, Pika Art (Pika Labs), Haiper, Vidu Studio, Sora, PixVerse, Krea, Veo, Seedance (ByteDance), Video Ocean, Seaweed, Stable Video Diffusion, Hunyuan (Tencent), Wanx (2.1), and more. It supports creation methods such as text-to-video, image-to-video, video style transfer, and character consistency animation, and offers over 150 video and image effects. It also comes with AI tools for video enhancement, quality improvement, face editing, and more. In use, no programming or design experience is required; simply upload an image or enter a text prompt to generate with one click. It relies on a cloud AI engine and multi-model parallel processing to achieve rapid generation. The platform is designed for browsers, with a smooth desktop experience and a responsive mobile web interface, requiring no app installation. For specific fees, pricing, and registration details, please refer to the official website.

Read more

Mochi-Full-Finetuner

AI video

Mochi-Full-Finetuner is an open-source tool repository for full fine-tuning or partial fine-tuning (e.g., specific DiT blocks) of the Mochi and Pusa 0.5 video generation models. It supports single-node and multi-node distributed training, and can use pre-encoded datasets or custom data. According to the official instructions, full fine-tuning is recommended to use 8 H800 or higher GPUs, and supports video lengths of up to 163 frames (about 5.4 seconds, 30 FPS). The tool provides complete environment installation steps, Mochi 1 weight download, dataset preparation, single-node/multi-node training scripts, and a method for converting fine-tuned checkpoints to safetensors, making it easy to replace the fine-tuned DiT weights into the original Mochi or Pusa model for video generation. In addition, the repository also cites related references for LoRA fine-tuning, but the current version does not support more memory-efficient training methods such as context parallelism, sequence parallelism, or tensor parallelism.

Read more

Mojo

AI video

Mojo is an AI-based product visual enhancement tool, officially described as designed to help users easily improve product presentation. Users only need to upload a product photo and select a preset effect to add dramatic effects such as smoke and glitter to a single image, and generate eye-catching videos within minutes. The tool focuses on fast, low-barrier product content creation, aimed at users who need to produce visual materials for product promotion or social sharing, and it can complete the process from image to video without complex operations. At present, the official information provided only includes the above functional overview and does not provide specific pricing plans, supported platforms, languages, export formats, usage restrictions, or detailed operating steps.

Read more

Monty

AI video

Monty is an AI short video production tool that automatically edits a single raw clip of you talking to the camera into a publishable short video (Reel/Shorts). It cuts the highlights, adds music, B-roll, and subtitles, aligns them to the beat, and then automatically publishes to YouTube, Instagram, TikTok, and Telegram. Users only need to provide one talking-head clip, with no need to re-shoot. Monty can also write scripts imitating the user's style, and supports automatic subtitles, multi-platform scheduled publishing, multilingual translation, and subtitle rebuilding. A free plan is offered, and it can be used without downloading anything.

Read more

Mora

AI video

Mora is an open-source multi-agent framework for general-purpose video generation. Through the collaboration of multiple visual AI agents, it completes tasks such as text-to-video generation, image-conditioned video generation, video extension, video editing, video connection, and simulating digital worlds. The project aims to reproduce and extend the capabilities of OpenAI Sora. According to official materials, Mora is already able to generate videos about 80 seconds long, but there is still a clear gap compared with Sora in aspects such as resolution, object consistency, and motion smoothness. The code has been released, but the training code and v2 paper have not yet been provided.

Read more

Motion by Mosaic

AI video

Motion by Mosaic is an AI agent tool for motion design. Users only need to describe their video requirements in one sentence, and it will automatically complete design, animation, and rendering, producing product launch videos, product demos, logo animations, social ads, and more. It can be used in the browser and can also be connected to Claude, ChatGPT, and any agent that supports MCP. The official description states that it shortens work that originally took hours into starting from a single prompt, with a single agent completing the entire process from idea to finished video, including research, design, motion, voiceover, and editing.

Read more

Motion Control

AI video

Motion Control (motioncontrol.app) is an AI image-to-video tool based on Kling motion control technology. Users upload a reference image and a motion reference video to transfer the movements, gestures, or dance from the reference video onto the static image, generating cinematic videos with frame-level motion precision, controllable camera movement, and stable character consistency. The tool supports the Kling Motion Control 2.6 and 3.0 models and can be paired with text prompts to control clothing, scenes, and visual style; the official statement says generation can be completed within minutes. Images support png/jpg/jpeg formats, and videos support MP4, with a duration of 3-30 seconds and a maximum size of 100MB. Applicable scenarios include character animation, dance motion transfer, full-body movement and facial expression synchronization, complex motion generation, camera planning previews, and product advertising proof of concept, among others. The site provides subscription plans, credit quotas, video quantity limits, and FAQs; specific pricing information is not currently provided.

Read more

Motion Control AI

AI video

Motion Control AI is an online AI video generation platform based on KlingAI's Kling 3.0 video model. By combining a reference image with a reference video, it transforms static images into dynamic videos and precisely replicates the motions, poses, facial expressions, and movement trajectories from the reference video. Users can upload a reference image and a motion video (supporting .mp4/.mov, with duration depending on character orientation, up to 30 seconds), and can choose whether to retain the original audio, while using text prompts to customize the scene, atmosphere, and motion effects. The platform emphasizes character consistency, realistic lighting and physics, complex motion control, cinematic camera movement, and precise trajectory control, outputting high-resolution, watermark-free MP4 videos. It is suitable for single-person realistic motion videos. Officially, it is recommended to match full-body or half-body framing, choose reference videos with large motion amplitude and moderate speed, and ensure sufficient space to avoid cropping. Generation time depends on motion complexity and server load; officially, a standard 5-second clip is usually completed within 3 to 5 minutes.

Read more

MotionControlAI

AI video

MotionControlAI is a motion control platform for cinematic-grade AI video creation. By mapping the motion, facial expressions, and camera behavior from a driving video onto a reference image, it generates finished shots with consistent characters, nuanced expressions, and controllable camera movement. Its workflow consists of four steps: select a high-quality reference frame, upload a driving motion video, enable element binding and enter camera language prompts, and after rendering, calibrate item by item and iterate. The platform emphasizes capabilities such as stable character identity, conveyance of micro-expressions and blinks, dynamic body and hand interaction, and clothing and spatial tracking. It also provides preset camera presets such as push-in, pull-out, pan, and tilt, and supports parameter archiving to enable repeatable production-grade iteration. The official description states that it can be used for free.

Read more

MovArt AI

AI video

MovArt is an all-in-one AI video generator that supports text-to-video, image-to-video, and video editing. It aggregates multiple flagship video models such as Veo 3.1, Sora 2, Wan 2.7, Kling, Hailuo, and Seedance in a single workspace, and also offers Nano Banana Pro image generation, capable of outputting broadcast-quality video up to 4K. It emphasizes multi-model switching, rapid iteration, and one-click viral effects. The platform also integrates tools such as a video editor, image editor, and background removal, making it suitable for scenarios including marketing and advertising, social media content, e-commerce products, education and training, film and television previsualization, and real estate and architecture. The official claims say it can be tried for free and requires no login.

Read more

movavi

AI video

Movavi is a suite of photo and video editing software aimed at ordinary users, focusing on speed, ease of use, and security, and the company states it has more than 70 million users worldwide. Its product line includes a video editor, photo editor, video converter, and screen recorder, and it offers the Movavi Unlimited bundle that integrates all programs. Its video editor provides basic functions such as cutting, cropping, color adjustment, and speed changes, as well as AI features such as one-click automatic subtitles, AI noise reduction, AI motion tracking, automatic removal of silent segments, and adjustable effects, and it also includes hundreds of filters, transitions, titles, and overlay assets. The new Movavi Templates is composed of pre-edited Blocks to form modular templates, allowing segments to be replaced and structure and text to be adjusted. The photo editor offers AI retouching, background replacement, object removal, and color correction presets. The video converter supports fast format conversion, video compression, and up to 8x AI video upscaling. The screen recorder supports full-screen or custom-area recording, webcam capture, and real-time annotation. The products offer a free trial version, but exports have a watermark, time limits, and some advanced features are locked; full functionality requires purchasing a lifetime license or annual subscription.

Read more

Muse AI

AI video

Muse AI is a browser-based creative workspace for editing photos, generating variant styles from reference images, and turning selected frames into short video ideas. Users can upload photos or start with a prompt, adjust background, lighting, style, and composition through natural language, and can choose image-to-image or text-to-image generation, then use a selected still frame as the starting frame to plan camera movement, pacing, and scene changes to generate a short video. The tool supports coordination of multiple reference images, creation history, and downloads, and offers annual subscription plans (Starter, Pro, Studio); specific feature usage is subject to credit and video quantity limits, and commercial use must comply with the relevant terms.

Read more

MuseV

AI video

MuseV is an open-source virtual human video generation framework by Tencent Music Entertainment's Lyra Lab, based on diffusion models and a "Visual Conditioned Parallel Denoising" approach, supporting infinite-length, high-fidelity video generation. It is compatible with the Stable Diffusion ecosystem, allowing the use of base_model, LoRA, ControlNet, etc.; it supports three types of tasks: Image2Video, Text2Image2Video, and Video2Video, as well as multiple reference image techniques such as IPAdapter, ReferenceOnly, ReferenceNet, and IPAdapterFaceID. The project provides virtual human video generation checkpoints trained on human datasets, and includes a Gradio interface for convenient local generation. The official release also includes the real-time lip-sync model MuseTalk and the pose-controlled image-to-video framework MusePose, which can be combined into an end-to-end digital human solution. The code is open-sourced under the MIT license, and the models are restricted to non-commercial research use only.

Read more