For Tripo, Meshy and Mixamo models · from $1
Tripo, Meshy and marketplace models come rigged for the body and dead from the neck up — no visemes, no blinks, nothing to drive. Upload your GLB, trace the lips and click the eyes, and about a minute later you have a validated VRM 1.0 that talks, blinks and emotes. You watch a video of every expression before you spend a dollar on it.
Get early access
GLB or VRM in · VRM 1.0 out · first model free · we'll email you when your invite is ready, plus the occasional product update.
How it works
Drop in a .glb from Tripo, Meshy, Mixamo or your own pipeline — or an existing .vrm to re-rig. Riggle normalises the scale, fixes the facing, bakes an A-pose to T-pose and renders your model's face straight on.
Outline the mouth, drop a point inside it, tag the upper and lower lip and click each eye centre. That's the whole job. You know where your model's lips are — pointing at them takes twenty seconds and gets a result guessing can't.
Your outline becomes a real lip mask on the mesh: Riggle flood-fills it across the surface, finds the cavity rim and zips the lips shut into a proper closed seam. The visemes are then genuine re-opens of your model's own mouth geometry — not a Gaussian bulge in the middle of the face.
You get a video of every expression playing back — aa, ih, ee, oh, ou, blink, left and right wink, happy, angry, surprised — plus a head turn. Like it? Download a validated VRM 1.0. Don't? Re-mark and run it again for free until you do.
Why Riggle
You spend twenty seconds. Mapping the skeleton, rebuilding the mouth, rendering the preview and validating the file are all automatic.
Riggle welds the mesh across import seams, flood-fills your traced lip region and zips it shut along a seam biased toward the upper lip — so the jaw does the closing, like a real mouth. Rest pose is a closed mouth; the visemes open it.
Models generated with an open mouth carry real cavity geometry — teeth, tongue, an inner rim. That's the asset most pipelines throw away, and it's what makes Riggle's visemes look like speech instead of a stretched texture.
Every expression is rendered and encoded into a short video with the same three-vrm stack your app will use, so what you watch is what you ship. Each shape is pixel-checked against the neutral face to catch a dead morph before you pay.
Tokens are spent on downloads, not attempts. Trace the lips again, nudge the eye centres, run it once more — you only pay when you take the model home.
Riggle vs. the alternatives
VRoid tools only touch VRoid topology. FACEIT and Polywink stop at FBX. A freelancer does great work in a day for the price of a month of Riggle.
| HANA_Tool | FACEIT | Polywink | Fiverr rigger | Riggle | |
|---|---|---|---|---|---|
| Works on AI-generated / custom topology | — | ✓ | ✓ | ✓ | ✓ |
| Outputs a ready-to-stream VRM 1.0 | ✓ | — | — | ✓ | ✓ |
| No Blender, Unity or plugins | — | — | — | ✓ | ✓ |
| Video preview + automated validation | — | — | — | — | ✓ |
| Your effort per model | One click | Hours in Blender | None | None | ~20 seconds |
| Turnaround | Instant | DIY | ~24h | ~1 day | About a minute |
| Price | Free · VRoid only | $48–130 | Studio quote | $10–100/model | From $1/model |
HANA_Tool is excellent — for VRoid models. The moment your model comes from an AI generator or a marketplace, the ecosystem sends you to Blender. That's the gap Riggle fills, and the twenty seconds of clicking is the price of not going there.
Pricing
Every account starts with enough tokens to rig a model for free. Marking, previewing and re-running cost nothing — a download costs 10 tokens, about $1 a model.
$0
Rig a model start to finish and see the real output.
Get early access
GLB or VRM in · VRM 1.0 out · first model free · we'll email you when your invite is ready, plus the occasional product update.
$9.99/ month
For creators shipping avatars regularly.
Get early access
GLB or VRM in · VRM 1.0 out · first model free · we'll email you when your invite is ready, plus the occasional product update.
$99/ year
Two months free, same everything.
Get early access
GLB or VRM in · VRM 1.0 out · first model free · we'll email you when your invite is ready, plus the occasional product update.
Prices in USD. Prefer no subscription? A one-time 100-token pack ($14.99) covers 10 downloads and stays valid for a year. Downloads cost 10 tokens each; marking, previews and re-runs are always free.
Riggle asks for twenty seconds of clicking instead of pretending to be one-click, and here's why: on continuous geometry no automatic falloff can tell a lower lip from a chin — we tried four approaches and the whole lower face smeared every time. Your traced lip line fixes that in one step, and everything after it is automatic. Results are best when the model was generated with its mouth open, since the cavity geometry is what the visemes re-open; a sealed, painted-on mouth has less to work with and Riggle will carve one instead. You watch the video preview before spending a token, so you always see the real result first.
Riggle is opening to a first group of creators. Leave your email and we'll send you an invite the day sign-ups open — your first model is free.
Get early access
GLB or VRM in · VRM 1.0 out · first model free · we'll email you when your invite is ready, plus the occasional product update.
Everything except the marking is automatic. The marking exists because automatic mouth detection genuinely doesn't work on arbitrary meshes: on continuous geometry there is no reliable way to tell a lower lip from a chin, so procedural approaches drag the whole jaw and cheeks along. Twenty seconds of your clicks turns a mediocre result into a good one, so we ask.
A VRM 1.0 file with the five visemes (aa, ih, ee, oh, ou), blink plus independent left and right winks, happy, angry and surprised expressions, and eye-bone look-at when your model has eye bones. Every shape is rendered and pixel-checked against the neutral face before delivery, and you get the preview video and stills alongside the model.
Any rigged humanoid .glb — Tripo, Meshy, Mixamo or your own export — plus existing .vrm files you want re-rigged. The body rig has to be there already; Riggle maps it to the VRM humanoid skeleton, rescales the model to human height and bakes an A-pose to T-pose automatically.
It works best on models generated with an open mouth, because the cavity geometry — teeth, tongue, inner lip rim — is what the visemes re-open. For flat or sealed mouths Riggle can carve a cavity from your outline, but the result is coarser. If you control the generation prompt, ask for an open mouth.
Re-mark it. The video preview comes before payment, and tokens are only spent when you download. Trace the lips differently or move the eye points and run it again as many times as you want — re-runs are free.
No. Everything runs in your browser: upload, mark, watch, download. Blender does run on our servers to do the actual work, but you never see it.
No, and we'd rather say so plainly. Riggle produces the VRM 1.0 expression set — visemes, blinks, emotions and look-at — which is what VRM apps drive for speech and reactions. Full 52-shape ARKit perfect sync is a different, larger job and it's on our roadmap, not in the product today.
Plan tokens refill every month and don't roll over. One-time pack tokens stay valid for a year. Downloading a model you already paid for is always free, even months later.