Bring an AI companion into your room
On camera, for content or just for company. This is the classic moment of having an AR companion in your room — a character that shares your actual space. See AI Companion in Your Room.
MODES
Summon: Call a Virtual Character into Your Real Room
Gesture at empty space on your desk. A character appears — and it notices you're there.
What it does
Most AR character apps need a pre-built 3D model, a compatible headset, or a tracked marker to work. LiveGen's Summon mode skips the 3D pipeline entirely: point your camera at a real space, make a gesture, and a virtual character appears in that space on your live video feed — aware of where you are and how you're moving, without any 3D modeling on your end.
Before: an empty desk, an empty room, a plain camera feed. After: the same real space with a virtual character present in it, reacting to your hand position, your gestures, and your movement — not a flat sticker pasted over the video, but a presence that responds to the room in real time.
This works entirely through 2D video, not 3D asset creation — so there's no rigging, no headset, and no app to install to see a character show up in your actual room. If you have ever wanted to summon a virtual character in AR without touching a game engine, this is the shortcut: a browser tab, a camera, and a gesture.
How it works
Summon runs on LiveGen's real-time engine in a mode built to place a character into a real scene and keep it responsive to you. There is no upload-render-download cycle. Your camera feed streams to the model over WebRTC, the model composites the summoned character into each frame as it senses your gestures and position, and the finished frame renders straight back into a standard <video> element in your browser. What you see on screen is the live result, frame after frame.
Open your camera
Point it at the real space you want the character to appear in. A clear, well-lit patch of desk, floor, or wall works best.
Gesture to summon
A hand gesture or on-screen action brings the character into frame, placed in the real space in front of you.
Interact live
The character tracks your gestures and position, responding in real time. When you like the moment, share the clip instantly with its own link.
Because the whole loop happens live, you direct the scene the way you would direct a person: move your hand, change your position, and the character responds to what you actually did — not to a prompt you submitted a minute ago.
What you can do with it
Create AR-style content for social video
Without any 3D modeling or game engine — a gesture-controlled AR character that reacts on camera makes for a scroll-stopping short clip.
Stage a virtual cosplay or collab scene
Where a summoned character appears alongside you in frame. See Virtual Cosplay.
Meet a character you already love
Fandom has a whole vocabulary for this: your oshi (the one you support), your comfort character, or, if you practice self-shipping the way a yumejoshi does, your F/O. The matching walkthroughs are a comfort character on your desk and meeting your oshi.
React and interact live
The character responds to your gestures as you make them, not on a delay, so reaction videos and duets feel genuinely interactive.
Follow a full walkthrough
If you want it broken down step by step. See How to Put a Virtual Character in Your Room.
Tips for the best results
Light the real space evenly
The model composites the character into your actual environment, so consistent lighting on the surface where the character appears helps it sit in the scene convincingly.
Keep a clear area for the character
Aim the camera at an uncluttered patch of desk, floor, or wall so there's room for the character to be placed and to react to your gestures.
Make deliberate gestures
Summon senses gestures and hand position, so slower, clearer movements read better than fast, tiny ones — especially the first time you call the character into frame.
Hold the camera steady
A stable phone or webcam gives the model a consistent scene to work with; a tripod or a propped-up phone beats hand-holding for longer takes.
Bring your own reference
Start from the character library, or upload a reference image to summon a custom character that matches a look you already have in mind.
Free exports are watermarked
You can start free and experiment as much as you want; removing the watermark and exporting in HD is part of the paid tiers (verify current pricing).
Real-time, not render-and-wait
Summon is a live transform, not a batch video generator. You are not typing a description, submitting a job, and waiting for a file to render. Your camera is the input, the model works on the stream as it arrives, and the character appears and reacts inside your live feed. That difference is the whole point: you can improvise, adjust, and perform, because the character is present with you in the moment rather than baked into a clip you receive later. If you want the underlying concept, see real-time interactive video.
Because it is billed by generation time — 1 credit = 1 second — you only spend on the seconds you actually run, and new users get free credits to try it.
Why LiveGen
Real-time, gesture-aware presence
The character reacts to your actual movement live — it's not a static overlay that ignores what you're doing. It's a virtual character live in AR, responding as you go.
Runs in your browser
No headset, no AR SDK to install, no game engine — just a camera and a browser tab, on desktop or mobile.
No 3D pipeline
No rigging, no modeling, no marker to print. Summon works through 2D live video, so a character can appear in your real room in seconds.
Instantly shareable
The moment a character appears in your room, the clip is ready to share with its own link.
One model, six modes
Summon sits alongside Live Morph, Bring to Life, and Freestyle — so once you're in, you can mix live effects without learning a new tool.
FAQ
Frequently asked questions
Do I need an AR headset or special hardware?
No. Summon runs through your regular camera and browser — no headset or dedicated AR hardware required.
Does the character actually respond to my gestures, or just appear once?
It tracks your gestures and position continuously, so it responds to your movement in real time rather than playing a fixed animation once.
Do I need any 3D modeling experience to use a character?
No. Summon works through 2D live video — there's no rigging, modeling, or game-engine setup involved.
Can I use this on a phone?
Yes — it runs in a mobile browser, so you can summon a character using your phone's camera.
What can I summon — only preset characters, or my own designs?
You can start from LiveGen's character library or bring your own reference image to summon a custom character.
How is Summon different from a face filter or a sticker?
A filter restyles your look and a sticker floats over the video. Summon places a separate character into the real space around you and keeps it aware of your gestures and position, so it behaves like a presence in the room rather than an overlay on your face.
Can other people see the character during a video call or live stream?
Summon renders the character into your live video feed, so whatever is watching that feed sees the composited result. It's built for real-time content and sharing, not as a private overlay only you can see.
Does it work in any room, or do I need a special background?
It works in ordinary rooms — no green screen or marker needed. An evenly lit, uncluttered spot for the character to appear in gives the most convincing result.
Play the world. In real time.
Open your camera and play any character live — free to start.
Start generating free