HANA_Tool won't touch your model? It only speaks VRoid.
HANA_Tool is the free, one-click answer for VRoid models, and it does full perfect sync — which Riggle doesn't. The catch is right there in the description: VRoid topology only.
HANA_Tool (hinzka's 52blendshapes-for-VRoid-face) adds all 52 ARKit perfect-sync shapes to VRoid models in one click, for free, and nothing beats it in that lane. But it depends on VRoid's exact face topology, so it can't touch an AI-generated, commissioned or marketplace model. Riggle covers that gap with a deliberately different deliverable: upload any rigged .glb or .vrm, mark the lips and eyes yourself in about twenty seconds, and get back a VRM 1.0 with working visemes (aa, ih, ee, oh, ou), blink plus independent blinkLeft and blinkRight, happy, angry, surprised and eye-bone lookAt — not the 52-shape ARKit set. VRoid model and you want perfect sync? HANA_Tool. AI-generated or custom model with a completely dead face? Riggle.
HANA_Tool: brilliant, but VRoid-only
HANA_Tool closes the VRoid segment cleanly. If you exported your avatar from VRoid Studio, it adds the 52 ARKit shapes and wires them in a single click, at no cost. There is no reason to pay anyone for perfect sync on a VRoid model.
The limitation is structural, not a missing feature: it works by knowing VRoid's exact face topology. Point it at a model built in Blender, generated by Tripo or Meshy, or bought on BOOTH or Gumroad, and there is no matching topology to work from. The whole VTuber ecosystem hits this wall, and the usual advice from there is 'go learn Blender'.
Riggle: any rigged topology, twenty seconds of your input
Riggle is built for exactly the models HANA_Tool can't read — with an honest trade in scope:
- Any rigged topology. There's no template to match. Your traced lip outline defines the mouth, so an AI-generated or marketplace mesh works the same way a clean one does. The body rig does have to exist already — Riggle maps skeletons onto the VRM humanoid bones, it doesn't create them.
- A mouth that really closes. Riggle welds the mesh across glTF import seams, flood-fills your outline across the surface, finds the mouth-cavity rim and zips the lips shut into a closed rest pose biased toward the upper lip — so the jaw does the closing and the visemes are re-opens of your model's own geometry.
- What you actually get. aa, ih, ee, oh, ou, blink with independent blinkLeft and blinkRight, happy, angry, surprised, and eye-bone lookAt when the rig has eye bones, exported as VRM 1.0. That is the VRM expression set — not the 52 ARKit shapes.
- See it move before you commit. Every expression is rendered into a preview video and pixel-diffed against the neutral render to catch a dead morph. Re-mark and re-run as often as you like; tokens are only spent on downloads.
Side by side
| HANA_Tool | Riggle | |
|---|---|---|
| All 52 ARKit perfect-sync shapes | Yes (one click) | No — roadmap, not product |
| Visemes, blink, emotions, look-at | Yes | Yes (VRM 1.0) |
| Works on non-VRoid topology | No | Yes — any rigged humanoid |
| Output | VRM (VRoid model) | VRM 1.0 + preview video |
| Your effort per model | One click | ~20 seconds of marking |
| Price | Free | Free tier (one model / month) · $9.99/mo for 10 |
Which should you use?
It isn't really a competition — they cover different models and deliver different things. If your avatar came straight out of VRoid Studio and you want full iPhone tracking, HANA_Tool covers that case, on VRoid topology only. If your model is AI-generated, commissioned, or a non-VRoid marketplace base, HANA_Tool has nothing to grip and Riggle does — as long as what you need is an avatar that lip-syncs, blinks and reacts rather than a full 52-shape rig. Plenty of creators end up using both: HANA_Tool on their VRoid avatar, Riggle on the Tripo model they generated last week.
Rig the model HANA_Tool can't read
If your model isn't VRoid, mark its lips and eyes in about twenty seconds and get back a VRM 1.0 that lip-syncs, blinks and emotes — with a video of every expression first. First model 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.
FAQ
Does HANA_Tool work on non-VRoid models?
No. HANA_Tool is built around VRoid's face topology, so it only serves VRoid-native models. Riggle works on any rigged humanoid mesh, but delivers the VRM 1.0 expression set — visemes, blinks, emotions, look-at — rather than the 52 ARKit shapes.
Is Riggle free like HANA_Tool?
There's a free tier: 10 tokens a month, which is one complete model. Marking, previewing and re-running never cost tokens — only a download does, at 10 tokens. Pro Monthly is $9.99 for 100 tokens a month (10 downloads), Pro Yearly is $99, and a one-time 100-token pack is $14.99 and stays valid for a year.
Can I use both?
Yes, on different models — HANA_Tool for your VRoid avatar and its perfect-sync rig, Riggle for the AI-generated or marketplace model HANA_Tool can't read.