Spine has been the default 2D skeletal animation tool in games for over a decade, and it has earned that. Exode added bones, IK and weight painting in the last year. On paper the feature lists overlap. In practice they are aimed at different artefacts, and choosing badly costs months.
Let me be concrete about where the line is.
What Spine is genuinely best at
Mesh deformation. Spine's free-form deformation and weighted meshes let an artist bend a shape in ways a bone hierarchy alone cannot. For characters — capes, tentacles, squash-and-stretch on a body — this is the difference between "rigged" and "alive".
Curve timelines. Spine's graph editor is built for animators who think in curves and holds. It is more expressive than a keyframe list with four easing presets.
Game engine pipelines. Official runtimes for Unity, Unreal, Godot, Cocos and more, with a decade of production hardening. If you ship a game, this is not a detail; it is the whole integration.
Attachments and skins. Swapping weapons, costumes, expressions at runtime is a first-class workflow.
A mature export format. .spine/.skel with versioned runtime compatibility, and tooling that game studios already have in their build pipeline.
If your artefact is a game character, stop here. Buy Spine. This post is not going to talk you out of the correct choice.
What Exode is doing with bones
Exode's bone system exists for a different reason: interface elements that need to bend.
That sounds narrow until you list the cases:
- an onboarding mascot that has to sit inside a responsive layout
- a brand illustration that deforms on scroll, next to real text
- an empty-state character that reacts to a form's validation state
- a loading skeleton with a squash-and-stretch envelope
- a hero object with a follow-path constraint and a gestural spin
In all of those, the rig is attached to an interface that also has flexbox, state, accessibility and data. Putting the rig in Spine means putting the whole interface in a canvas player, and then you have lost the interface.
Exode's bet is that a good enough rig inside a real layout beats a great rig outside one, for this class of work.
Feature comparison, without the marketing
| Capability | Spine | Exode |
|---|---|---|
| Bone hierarchy, master/child | Yes, deep | Yes |
| Inverse kinematics | Yes, mature | Yes |
| Weight painting / vertex weights | Yes, batch and per-vertex | Yes — batch weighting and a live weights table shipped |
| Free-form mesh deformation | Yes, best in class | Limited |
| Curve editor | Yes, animator-grade | Yes, with a shared ease creator |
| Physics constraints | Yes | Yes — rigidbodies, damping, restitution |
| Flexbox / responsive layout | No | Yes |
| Text layout and typography | Minimal | Full |
| State machines for app logic | No (animation states only) | Yes — typed inputs, listeners, transitions |
| DOM semantics / accessibility | No (canvas) | Yes (compiled components) |
| Game engine runtimes | Excellent | Web/app focus |
| Export |
.skel / JSON for engines |
React, Vue, Svelte, Solid, Angular, Swift, Kotlin, RN |
| Format | Binary-ish, tool-owned |
.exode JSON with schema |
The honest gaps
The Exode rigging audit is public in the repo, and it is worth reading because it lists the eight gaps the team closed rather than pretending they never existed: destructive auto-bind on selection, single-vertex selection, a read-only weights table, clicking a bone exiting edit mode, static weight visualisation, destructive re-binding, no independent Bézier handle weighting, and rigid normalisation.
They are closed now. That does not make the tool equal to Spine for character animation. Mesh deformation is still shallower, and the curve tooling is younger. If your art director wants to spend three days on a cape, Spine is still where that happens.
A workflow that uses both
This is the part most comparisons miss: you do not have to choose.
A pattern that works:
- Rig and animate the character in Spine.
- Export a baked sequence or a sprite sheet for the states that are decorative.
- Keep the interface-level rigs — mascot placement, UI deformation, gestures — in Exode, where the layout and state live.
- Drive both from the same state machine so the character's expression and the UI's state cannot disagree.
It is more moving parts than one tool. It is also how a lot of shipped products actually work, because no single tool is best at both drawing a character and laying out a checkout.
How to decide in ten minutes
Ask three questions.
1. Is the artefact a character or an interface?
Character → Spine. Interface with a rigged element → Exode. A game with a UI → Spine for the art, something DOM-based for the UI.
2. Does the rig need to sit inside responsive layout?
If yes, a canvas player will fight you on every breakpoint. That is the strongest signal for an interface-native rig.
3. Who maintains it after launch?
If the animation team owns the pipeline and ships through Unity, Spine's tooling is the path of least resistance. If a front-end team owns it and the rig lives in a React repo, a JSON format they can review is worth more than a deeper mesh tool they cannot open.
The thing I would tell a junior
Rigging is a craft, and the tool matters less than the rig. A clean hierarchy with sensible weights in a simpler tool will beat a messy rig in a powerful one every time. Learn the craft on whichever tool your team already has, and switch only when a specific limitation blocks you — not because a comparison table has more green ticks.
UX Baba writes about interaction design and tooling. Exode's bone and weights work is documented publicly in the repo if you want the unglamorous details. The editor is at exodeui.com.
Top comments (0)