Real-time character swap
CharX
Re-render a person as a chosen character while preserving their live motion, gaze, and expression.
Explore the real-time interactive AI video model built to transform characters, clothing, scenes, and motion as the camera moves.
Xmax X2.0 capabilities
Explore five ways to reshape a live scene: CharX, ClothX, VibeX, MoX, and DimX.
Product imagery is credited to Xmax AI. Read our source notes.
Showing CharX, slide 1 of 5.
Official API workflow
Follow the flow provided by the official Xmax AI platform. XmaxAI.org is an independent guide and does not operate the model.
Start with a camera, uploaded video, or reference image.
Restyle the scene as a luminous cinematic world...
Pick CharX, ClothX, VibeX, MoX, DimX, or the open prompt mode.
Stream the source, update the prompt or reference, and end the session when finished.
Where interactive video fits
Xmax X2.0 is aimed at moments where a generated video must react while someone is still watching, moving, or touching the screen.
Let audience input change a host’s character, outfit, or entire visual world while the stream is live.
Give shoppers a moving, camera-based preview instead of a single static clothing composite.
Explore a future where streets, buildings, and nearby objects can be re-rendered in real time.
Turn IP, collectibles, and virtual companions into responsive presences inside a real camera scene.
Technical snapshot
Public descriptions say X2.0 generates frame by frame with an autoregressive streaming architecture, then uses distillation, sparsity, quantization, and persistent context to keep latency and long-session drift under control.
960p
up to 24fps
Reported cloud/consumer-GPU output
384p
mobile demo
Reported local iPhone inference
5 + Free
interaction modes
CharX, ClothX, VibeX, MoX, DimX
API
available now
Official JavaScript SDK and live platform
These figures are company or media claims, not independent benchmarks. Mobile 384p and higher-resolution output refer to different hardware contexts.
Read the technical guideXmaxAI.org launch pricing
Independent XmaxAI.org early-access SaaS plans for workflow, media, and collaboration tools built around Xmax video production.
For solo creators building repeatable Xmax video workflows.
Billed monthly
What is included
For creators shipping higher-volume, client-ready work.
Billed monthly
What is included
For small teams producing campaigns and shared projects.
Billed monthly
What is included
XmaxAI.org pricing is for this independent SaaS workspace. Official Xmax X2.0 model and API usage is separate. The official global platform currently lists 3,600 credits for $6; at 1 credit per second, that equals $6 per generated hour.
Early-access plan allocations may change before checkout.
Independent product guide
Xmax AI is building a different category from the familiar text-to-video workflow. Instead of returning only a finished clip, Xmax X2.0 is designed to keep generating while a live source changes.
The practical difference is intervention. A person can move in front of a camera, enter a new prompt, drag a subject, or make a gesture and the visual result can respond during the same session.
The public X2.0 release groups this interaction model into five named modes: CharX, ClothX, VibeX, MoX, and DimX, plus a more open Free mode for custom instructions.
That makes the technology relevant to live commerce, streaming, virtual try-on, interactive entertainment, tourism displays, smart glasses, and responsive virtual characters.
| Dimension | Traditional AI video | Xmax X2.0 approach |
|---|---|---|
| Output | A finished video file | A continuously re-rendered video stream |
| Interaction | Prompt, wait, review | Prompt, camera, gesture, touch |
| Timing | Generation before playback | Generation during the experience |
| Typical use | Ads, films, social clips | Live video, try-on, AR, interactive IP |
Research library
Independent guides built from public product and launch materials. Last reviewed .
Xmax AI FAQ
Xmax AI is a team building real-time interactive AI video models. Its public product direction focuses on changing a live camera or video stream as a person moves, gestures, or enters a prompt.
Xmax X2.0 is the company’s second-generation interactive video model, announced on July 15, 2026. It groups real-time character, clothing, style, motion, and spatial interaction tools into one platform.
No. Xmax X2.0 is described as a video generation and re-rendering model. It can accept text instructions, but its primary output is an interactive video stream rather than a text conversation.
Join the XmaxAI.org early-access list to follow the in-site workspace as generation workflows, media tools, and future model integrations become available.
As of July 26, 2026, the official global platform lists 3,600 credits for $6. At one credit per second, that equals $6 per generated hour. Mainland China pricing is listed at ¥20 per hour. XmaxAI.org plans are independent SaaS workspace plans, and official model usage is billed separately.
GeekPark reports that the team demonstrated local iPhone inference at about 384p. The same report distinguishes that mobile demo from the higher 960p at 24fps output discussed for larger hardware.
No. xmaxai.org is an independent guide and SaaS project. It explains the public Xmax ecosystem while keeping its own research, workspace, pricing, and account experience separate.
XmaxAI.org early access
Join the in-site workspace waitlist for workflow presets, media organization, collaboration, and future model integrations.