Image to VRM — a PNG or JPG in, a talking VRM 1.0 out
A VRM is a 3D avatar with a skeleton and a face. A PNG is a flat picture. Getting from one to the other used to mean a 3D generator, an auto-rigger and an afternoon in Unity - or rebuilding the character by hand in VRoid Studio. Riggle does the whole chain from the picture.
To turn an image into a VRM, upload a PNG or JPG of one character - full body, facing forward, arms away from the body, plain background - to Riggle. You get back a VRM 1.0 of that character: a textured 3D body with a skeleton, the five visemes (aa, ih, ee, oh, ou), blink with independent left and right, happy, angry and surprised. You watch a video of every expression before you download. The 3D model takes about two minutes and 15 tokens from a Pro plan or a token pack, and the download is 10 tokens. No 3D software, VRoid Studio, Unity or UniVRM is involved.
What you get from the picture
- A 3D model. A textured 3D body of your character: the front is what you drew or generated, the back and sides are filled in to match.
- A skeleton. Riggle adds one so the body can move. If it needs your help, it asks you to check it on a front view.
- A face that moves. Visemes, blinks and emotions, made from the character's own mouth and eyes.
- A VRM 1.0. Ready for VRM 1.0 apps, with a video of every expression to watch before you download.
The picture, the 3D model and the VRM all stay on your account under Assets, each downloadable on its own.
Which pictures work
The 3D model can only have what the picture shows, so the picture decides almost everything. What works:
- One character, head to feet. The whole body inside the frame, with a small margin.
- Facing forward, arms away from the body. A T-pose or A-pose. Arms pressed to the sides fuse to the torso.
- A plain background. White or flat colour. Anything else in the frame - a floor, a shadow, a prop - can become part of the model.
- The mouth open, ideally. A character with its mouth open gives the most natural mouth shapes, and the model keeps whatever the picture shows - see why an open mouth decides your visemes.
What does not: a headshot or bust (a VRM needs legs), a group, a side or three-quarter view, a busy scene, or a four-legged creature, which VRM's humanoid skeleton has nowhere to put. Stylised and anime characters, generated or drawn, are the case it is built for.
Either PNG or JPG is fine, up to 100 MB.
Step by step
- Sign in and open Studio.
- Drop the PNG or JPG where you would drop a model. Riggle tells you the 3D model costs 15 tokens and how many you have; if you are short, the plans and the one-time token pack open on the same page and you come straight back to your picture.
- Confirm, and wait about two minutes while the model is built.
- Riggle shows you your model's face and asks you to point out the mouth and eyes - a few clicks.
- Watch the preview video of every expression. Run the face again as often as you like - that part is free.
- Download the VRM 1.0 for 10 tokens.
Image to VRM without VRoid Studio
The usual advice for turning a drawing into a VRM is VRoid Studio: open it, and rebuild the character with its sliders and hair tools, using the picture as a reference. That works when the character fits VRoid's anime template and you have an afternoon. It does not use the picture at all - it uses you.
Riggle builds the model from the picture itself, so the proportions, outfit and style are the ones you drew or generated, not the nearest VRoid preset. The trade is control: you get one model per picture, and if it is not what you wanted, the fix is a better picture. If you want to steer the 3D step yourself, build the model in a 3D generator and upload the GLB instead - rigged or not.
What it costs
Picture to 3D model costs 15 tokens, from Pro or a token pack, charged when the build starts and refunded if it fails. Downloading the finished VRM is 10 tokens, once per model. The previews and re-runs in between are free.
Pro is $4.99 a month or $49.99 a year, with 50 tokens a month - two avatars from pictures. A one-time 75-token pack ($9.99, valid for a year) tops it up.
Turn your picture into an avatar
Upload a PNG or JPG of your character and watch it talk, blink and react before you download the VRM 1.0.
Upload my pictureFAQ
Can I convert a PNG to VRM for free?
Not from the picture: building the 3D model costs 15 tokens, more than the free plan's 10 a month. What is free is the rest of the chain - if you get a 3D model of the character some other way, upload the .glb, rigged or not, and the free plan's 10 monthly tokens cover one VRM download.
Does it work with a JPG?
Yes. PNG and JPG (.jpg or .jpeg) are both accepted, up to 100 MB.
Can I use a photo of a real person?
Only of someone who agreed to it, and expect less: the 3D step is at its best on stylised characters with clean shapes and flat colours. The same framing rules apply - full body, facing forward, arms out, plain background.
Will a headshot or bust work?
No. A VRM is a full humanoid with legs, and the model can only have what the picture shows. Use a full-body picture of the character.
What if the 3D model comes out wrong?
If the build fails, the 15 tokens come back automatically. If it succeeds but is not the character you wanted, the fix is upstream: a cleaner picture, closer to the checklist above, gives a cleaner model.
Which apps can load the VRM?
Riggle exports VRM 1.0, which apps that read VRM 1.0 load as-is. Some VTuber software still reads only VRM 0.x and will refuse a 1.0 file, so check which version yours expects.
Do I own the avatar?
Riggle claims nothing over what you make. Commercial use is governed by the rights to the picture - your own drawing, or the terms of the image generator that made it.