One API. Every AI modality.

Seedance 2.5

Seedance 2.5 API is now available on Exgravity ai! It provides developers with ByteDance's latest video model. This model can generate native videos up to 30 seconds long at once, using text, single images, or up to 50 multimodal references, including synchronized audio and multilingual text within the frame. On Exgravity ai, you can access it with a single key, and its subject consistency and improved physical properties help maintain the coherence of long takes.

Wan 3.0

Wan 3.0 API is the next-generation product in Alibaba's Wan video series, designed to elevate long-form video generation, multi-reference control, and audio-video quality to new heights. Exgravity AI has already hosted Wan 2.7, 2.6, and 2.5, while Wan 3.0 runs with the same unified key and requires no additional setup. Start building now. Scroll down to the showcase area to see the creative capabilities of Wan 3.0.

MiniMax H3

MiniMax H3 is MiniMax's multimodal video series, supporting text, images, and reference content for guided creation. Among supported routes, it can retain the subject in reference media, provide flexible aspect ratios, and combine generated sound with visuals via H3 Developer; Output configuration is selected by End. Exgravity AI integrates the entire series through an OpenAI-compatible key and offers transparent pay-as-you-go pricing, with a standard rate of $0.038 per second. Start building now.

Seedance 2.0

Seedance 2.0 is ByteDance's video model for production environments, dedicated to precise shot creation. Convert prompts into video, generate animations using first frame images with optional tail frame guidance, or shape results by combining reference media and optional connected search. Exgravity AI integrates these workflows into a unified API, offering transparent pay-as-you-go pricing and an OpenAI-compatible key. Start building now.

Gemini Omni Flash

Gemini Omni is a native multimodal video series from Google DeepMind, supporting prompt-based creation and editing. You can generate videos from text, bring static images to life, or modify existing footage with optional visual references while preserving unchanged content. Flexible resolution, aspect ratio, and duration control to meet diverse production workflows. Exgravity AI provides unified access through a transparent, pay-as-you-go pricing model. Start building now.

Seedance 2.0 Mini

Seedance 2.0 Mini brings ByteDance's multimodal video generation technology into workflows with extremely high speed and cost requirements. It delivers Seedance 2.0's core capabilities in a lighter footprint—faster generation speed, lower cost per video—and integrates with your existing API. For teams running high-throughput pipelines or conducting large-scale prototyping, Mini is the most practical default choice.

Suno Chirp

The Suno Chirp API provides developers with a Suno music generation model on Exgravity AI through a single key. Using text prompts and optional lyrics, it can generate a complete song containing vocals and instruments, returning two versions per request, and offers both pure music and multilingual options. The product lineup includes Chirp V6, V6 Wild, and V6 Mini, charged per generation, starting at $0.132.

Wan 2.7

The Wan 2.7 API provides developers with Alibaba's one-stop video suite, covering text-to-video, image-to-video, reference-to-video, and video editing, plus image generation features. It can generate native 1080p video clips up to 15 seconds long with synchronized audio, supports frame control at the end of the frame, and supports up to 5 role references. (For more information on native 30-second videos and up to 20 mixed references, please refer to the latest [Wan 3.0 API page](https://www.atlascloud.ai/models/wan-3.0)。）

Seedream 4.7

ByteDance's Seedream 4.7 API unifies text-to-image generation, precise editing, and continuous image workflows within a single series. It can follow instructions while preserving the subject identity, lighting, and local structure in the source image, and can generate a coherent set of images in a single request. With an OpenAI-compatible Exgravity AI key, all four variants are available and transparent pay-as-you-go pricing is available. Start building now.

Vidu Q4 Preview

Vidu Q4 is Vidu's latest flagship video model, designed specifically for image-to-video and reference image-to-video generation. It can animate individual static images and automatically create scene transitions, providing two input paths for different product workflows. Exgravity AI provides dedicated API endpoints and releases parameter schemas at a verified standard price of $0.04/second. Start building now.

Hailuo 02

Hailuo 02 is Hailuo AI's video generation model series, designed for developers who need to flexibly balance performance across multiple levels. Its Standard model has proven to be 2.5 times more efficient, with an 85% response rate for complex instructions, capable of generating high-quality video. On Exgravity AI, billed at standard prices, Fast costs $0.10 per call, Standard $0.28, Pro $0.49, helping teams match suitable models for different workloads based on budget and quality requirements. Start building now.

PixVerse C1

Pixverse c1 is PixVerse's first AI video model specifically designed for film and television production. Its motion engine, cinematic visual effects, reference guidance consistency control, and multi-panel storyboard conversion capabilities support coherent and controllable multi-shot sequences. With Exgravity ai, developers can call the model at standard prices with just one OpenAI-compatible key, costing as little as $0.03 per second. Start building now.

Midjourney 8.2

Midjourney 8.2 brings the full range of Midjourney image and video creation capabilities to developers through Exgravity ai. Supports text generation, reconcepting or redrawing existing images, mixing 2 to 5 inputs, removing backgrounds, or turning a single image into 4 5-second 480p or 720p clips. Access all workflows with an OpenAI-compatible key and enjoy transparent pay-as-you-go pricing and reliable service availability. Start building now.

Happyhorse 1.1

HappyHorse 1.1 is Alibaba's developer video generation model line, supporting video generation from text, source frames, or visual references. It can convert prompts into video clips, animate specified first frames based on optional guidance, and also support precise visual control using 1 to 9 reference images. Exgravity AI integrates these workflows into a single access and offers transparent pay-as-you-go pricing. Start building now.

Gemini Omni 1.1 Flash

Gemini Omni 1.1 Flash is a native multimodal video series from Google DeepMind, designed for flexible production workflows. Edit existing footage with text commands, maintain visual consistency with up to 10 reference images and 3 video clips, or extend shots with satically connected clips for up to 40 seconds. Access all workflows on Exgravity AI with an OpenAI-compatible API key, pay-as-you-go, and support for Day-0 launch. Start building now.

Veo 3.1 Lite

Veo 3.1 Lite is a series of video models launched by Google for developers, designed for building scalable production lines. Depending on the endpoint used, you can choose flexible duration options to generate an 8-second transition video between a specified first and last frame, or generate video with a common aspect ratio. Exgravity AI delivers these capabilities through an OpenAI-compatible key, offering transparent pay-as-you-go pricing and streamlining API access processes. Start building now.

Veo 3.1 Fast

Veo 3.1 Fast is Google's focused speed video generation series, designed for responsive production workflows and rapid creative iteration. Its prompt and keyframe workflow delivers smooth, lively motion while maintaining excellent visual fidelity. Exgravity AI offers Day-0 access, unified keys, transparent pay-as-you-go pricing, and a verified standard base price of $0.08. Start building now.

Nano Banana 2.1

Nano Banana 2.1 is a series of image models built on Gemini 3.6 Flash by Google. It delivers Pro-level generation and editing capabilities at Flash-level speeds, supporting natural language editing and fusion of 1 to 14 input images, and maintaining consistency for up to 4 characters and 10 objects. With Exgravity ai, developers can access the series of models with just an OpenAI-compatible key, enjoying reliable uptime guarantees and Day-0 availability. Start building now.

Kling 2.0

Kling 2.0 is Kuaishou's video generation model series, supporting text-to-video and image-to-video creation. Its Master model combines high-fidelity visuals with fine lighting and realistic camera performance, providing developers with two focused workflows to help them convert prompts or source images into exquisite video clips. On Exgravity AI, developers only need an OpenAI-compatible key to access both models and enjoy Day-0 availability and a simple integration experience. Start building now.

Kling 2.1

Kling v2.1 is Kuaishou's AI video model series designed for developers, designed for production-grade video workflows. The series supports fast 720p draft generation, clear and smooth image animations, professional-grade visual depth, complex prompt word understanding, advanced motion logic, and enhanced dynamic camera rendering. Access these capabilities with a single OpenAI-compatible key via Exgravity AI's unified API, with Day-0 model availability. Start building now.

Kling 2.5 Turbo Pro

Kling v2.5 Turbo Pro is a video generation model launched by Kuaishou, capable of converting prompts and source images into high-quality video clips. It offers cinematic motion accuracy and enhanced temporal stability, helping scenes stay coherent even when the subject and camera move. Exgravity AI offers both generation modes through a unified API, using transparent pay-as-you-go pricing and equipped with an OpenAI-compatible key. Start building now.

Kling 2.6

Kling v2.6 is Kuaishou's video model series, supporting cinematic-grade text-to-video and image-to-video creation, AI digital humans, and motion migration based on reference content. It supports flexible aspect ratios, richer dynamic effects, stable subject identities, and cross-professional workflow time consistency. Exgravity AI offers Pro and Standard options through a unified endpoint and uses transparent pay-as-you-go pricing. Start building now.

Kling 3.0 Turbo

Kling V3.0 Turbo is Kuaishou's video generation series built on MVL technology, designed for developers who want to create dynamic cinematic video clips. Dedicated endpoints can convert creative input into generated video for each supported mode, providing a clear, focused workflow. On Exgravity AI, integration from prototype to production can be simplified using an OpenAI-compatible key, transparent pay-as-you-go pricing, and direct API access. Start building now.

Kling o1

Kling O1 is a video model series developed by Kuaishou based on multimodal visual language technology. Its text-to-video and image-to-video workflows combine language with visual context to generate coherent scene dynamics based on prompts or source images. Exgravity AI offers both modes through out-of-the-box REST APIs, with no cold start and transparent pay-as-you-go pricing. Start building now.

Seedream 5.0 Flash

Seedream v5.0 Flash is ByteDance's series of image models designed for image generation, editing, and structured asset extraction. Its layer decomposition endpoint can split an image into a base image and up to 16 sequentially arranged PNG layers, providing bounding boxes, names, descriptions, and z-index data. Exgravity AI offers pay-as-you-go access through three dedicated endpoints, with a verified standard base price of $0.018. Start building now.

Grok Imagine Video 1.5

Grok Imagine Video 1.5 is xAI's video generation series for developers, suitable for scenarios requiring controllable motion effects based on prompts and visual inputs. Use natural language motion commands to add animations to the initial frame, select 480p or 720p in the reference workflow, and add optional reference speech to guide the generation. Access supported modes with an OpenAI-compatible Exgravity AI key and enjoy transparent pay-as-you-go pricing. Start building now.

Grok Imagine Video

Grok Imagine Video gives developers access to the xAI video model family for creating, animating, extending, and editing video clips. Generates 1 to 15-second video via prompts, guides people, objects, or styles with up to 7 reference images, and supports 480p or 720p. Exgravity AI integrates these capabilities through OpenAI-compatible endpoints and offers transparent pay-as-you-go pricing. Start building now.

MiniMax H3 Max

MiniMax H3 Max from MiniMax can generate cinematic videos and supports optional end-frame guidance when animating source images, allowing developers to more precisely control the convergence effect of each shot. Exgravity ai provides pay-as-you-go access via an OpenAI-compatible API. The standard price for H3 Max is $0.05 per second, and Turbo is $0.025 per second. Both versions are charged by the second, with transparent pricing. Start building now.

MiniMax H3 Fast

MiniMax H3 Fast is MiniMax's prompt-driven video generation series, supporting cinematic text-to-video, controllable first-frame animation (optional end frame), and reference-image-based video clips that preserve the source subject. Through Exgravity ai's unified API, access the entire workflow series at a standard pay-as-you-go price of $0.046 per second, covering the full series with a single integration. Start building now.

Nano Banana Pro

Nano Banana Pro is Google's next-generation Nano Banana image model designed for production-grade visual workflows. It can generate sharper details, richer color control, and faster diffusion images, while dedicated editing endpoints support precise AI-driven adjustments. Through Exgravity ai, developers can choose Text-to-image, Edit, Ultra, or Developer endpoints, with standard prices as low as $0.14 per call. Start building now.

FLUX 3

FLUX 3 API brings Black Forest Labs' latest foundation model into your tech stack. This is a unified Self-Flow architecture trained jointly on images, videos, and audio. It can generate clips up to 20 seconds long, supporting multilingual dialogues, sound effects, and ambient sounds, or generate text-to-image content with accurate layout effects. Exgravity ai provides service via an OpenAI-compatible key, with pay-per-call pricing and support for Day-0 access.

VEED

Perfect for AI Avatar Video Creation

Youchuan

Youchuan API now covers V8.1 and V8.2 generations, including static images and dynamic videos. The prompt-driven mode returns four variants at a time, images rendered in native 2K HD without upscaling; hybrid mode can merge up to five references; a single static image can generate a 5-second video clip in 480p or 720p. Exgravity ai provides service via an OpenAI-compatible key, supports Day-0 access, and charges as little as $0.086 per call. Start building now.

Seedream 5.0 Lite

Seedream 5.0 Lite API brings ByteDance’s inference-priority image model to Exgravity ai. This is the first image model combining real-time web search with visual reasoning, making its generation process based on real-time information and real-world physics, with native editing support. On Exgravity ai, only one OpenAI-compatible key is needed, with single-image prices as low as $0.032 and support for sequences of up to 15 coherent frames.

Kling V3.0

Kling 3.0 API brings Kuaishou’s flagship video suite to Exgravity ai via an OpenAI-compatible key. It includes two models: Kling 3.0 for AI Director storytelling, multilingual lip-sync, and precise on-screen text; Kling 3.0 Omni (O3) for cloning subjects and audio through short videos or images. Both can generate native audio in a single process, supporting up to 4K output. Build cinematic storytelling, global marketing, multilingual advertising, and serialized character content on a reliable infrastructure.

Seedream 4.5

Seedream 4.5 API provides developers with ByteDance’s latest image model through a single endpoint. It combines 4K text-to-image generation with reference-based editing, accurately rendering typography and maintaining subject consistency across up to 10 reference images. Exgravity ai offers this service via an OpenAI-compatible API, with reliable uptime and transparent pricing ($0.04 per image), no subscription required.

Nano Banana

Nano Banana brings Google’s Gemini image generation and editing model to developers via Exgravity ai. You can create visual content from text or modify existing images, outputting in PNG or JPEG, and supporting square, portrait, landscape, and ultra-wide layouts. Optional Base64 output supports API workflows. Exgravity ai also provides an OpenAI-compatible interface, Day-0 access, and transparent pay-as-you-go pricing. Start building now.

Hailuo 2.3

Hailuo 2.3, developed by MiniMax, brings professional-grade video generation capabilities into the creative and production workflow. Standard and Pro provide advanced movement performance, realistic physics, cinematic visuals, detailed scene evolution, and coherent character portrayals based on reliable prompt follow-up capabilities; Fast supports rapid iteration. Exgravity AI offers all five variants with transparent, pay-as-you-go pricing. Start building now.

Wan 2.6

Wan 2.6 API constructs multi-camera stories with a single prompt word, rather than a single segment. Alibaba's narrative engine switches between shots with coherent transitions, while maintaining stable consistency in character identity, voice, and style throughout the scene. It covers text-to-video, image-to-video, and reference-to-video, outputs 1080p, with synchronized audio and multilingual lip-sync features. It can be used on Exgravity AI with a single key, alongside over 300 models.

Seedance 1.5 Pro

Seedance 1.5 Pro API brings ByteDance's audio-video model to developers through one endpoint. It generates video with synchronized native audio and multilingual lip sync, follows cinematic camera moves such as pans, orbits, and zooms, and keeps characters consistent across 4 to 12 second clips. Exgravity ai serves both text-to-video and image-to-video with reliable uptime and transparent pricing from $0.018 per second.

Ltx-2

The LTX-2 API brings Lightricks' open-weighted video series into production, enabling one-time video generation and synchronized audio. You can generate videos from text prompts, add animations to static images, or extend existing clips; Each result comes with matching lip-sync, foley, and ambient sounds. On Exgravity ai, you can access it with an OpenAI-compatible key, enjoying transparent pay-as-you-go pricing and Day-0 access. Start building now.

Nvidia Cosmos

Explore Nvidia Cosmos API on Exgravity ai.

Seedance

Seedance 1 Pro is ByteDance's video generation series designed for cinematic, multi-shot creation. It maintains scene coherence between clips, transforms static images into expressive dynamic images, and ensures consistent details between frames. On Exgravity AI, a single OpenAI-compatible key connects all v1 Pro interfaces; Billing by the second and convenient model switching support workflows ranging from rapid prototyping to production-level rendering.

Van

Van is a flagship AI video series built on the Wan 2.5 and 2.6 frameworks, with the Van API unifying its complete text-to-video and image-to-video product lines on Exgravity AI. Render up to 1080p cinematic footage, then select the speed-optimized 2.6 model for fast iteration or the cost-effective 2.5 variant for high-volume processing. Each model is released on Day-0 with a single OpenAI-compatible key, using transparent pay-as-you-go pricing. Start building now.

Wan 2.5

The Wan 2.5 API can generate video and audio at once, perfectly aligning voice, sound effects, and lip sync without extra steps. The model is built on Alibaba's Diffusion Transformer architecture, supporting video generation at 480p to 1080p resolution, 5-second or 10-second length, and reliable synchronization even for Chinese prompts. Access it and over 300 other models on Exgravity AI with a single key.

Kling o3

Kling O3 is a unified multimodal video model built on Kuaishou's MVL technology, supporting text, images, reference images, and video inputs. It guides movement through the first and last frames, maintains consistency of characters, props, or scenes with up to seven reference images, and allows you to choose Standard or Pro versions for each workflow. Use this model series on Exgravity AI to enjoy transparent pay-as-you-go pricing and reliable infrastructure. Start building now.

Wan

The Wan API introduces Alibaba's open-source Wan video model into Exgravity AI through a unified key. Wan 2.2 pioneered the Mixture-of-Experts architecture in video diffusion, enhancing model capacity and motion control capabilities while maintaining the same inference costs. It supports text-to-video, image-to-video, and video-to-video, featuring front-end frame control, video expansion, and image quality enhancement (super-resolution) functions, all accessible with over 300 models.

Tools

Open, advanced large-scale image generative models that power high-fidelity creation and editing with modular APIs, reproducible training, built-in safety guardrails, and elastic, production-grade inference at scale.