Blender says the bind worked. It says that either way.

You select the mesh and the armature, press Automatic Weights, and Blender reports success. On an ordinary humanoid it has just left four bones out of nineteen with no weight at all — both hands and both feet, every single time. On a coarser mesh, six. On a real production model it reported success having bound not one vertex out of eleven thousand seven hundred and sixty three.

Nothing in the interface tells you. You find out when you animate.

What it finds, and how it proves each one

A weight counter can tell you a group is empty. It cannot tell you a bone moves nothing, because the weight may sit on a group the modifier never reads. So BindProof does not count — it measures.

Bones that move nothing

Each suspect bone is turned 35 degrees and the moved vertices are counted: zero. Printed right beside it, a bone that does carry weight — 385 vertices, up to 102.9 mm. The zero is only worth something because the other number is there to show the measurement works.

Vertices bound to nothing

The rig is posed and the vertex is shown standing still while its mesh walks away: 0.00 mm against 460.39 mm. If the mesh has nowhere to go in this frame range, BindProof says nothing rather than printing "0.00 against 0.00" — an empty proof is worse than no proof.

A bind that did not happen at all

Reported on its own, not as a count of vertices. Eleven thousand unbound vertices is not a patchy skin; it is an operation that silently did nothing, and the two need different words.

More influences than your engine will read

Unity, Unreal and Godot take four influences per vertex and drop the rest without a word on import. BindProof counts them and works out from the skinning maths how far the worst vertex will move when that happens — so you know whether it matters on this mesh before you find out in the engine.

Weights that do not add up to one

Blender normalises on the fly for its own armature modifier, so you will never see it in the viewport — and Blender's own Automatic Weights leaves most of a mesh that way: 1325 vertices out of 1932 on a measured character. Said plainly, so you know it is normal inside Blender. It stops being normal on export, where the skin shrinks towards the rig or blows away from it, and BindProof tells you by how many millimetres.

Weights on a bone that has Deform switched off

The group exists, the modifier does not read it. From the outside it looks bound. It is a hole.

It does not cry wolf

A weight counter finds problems that are not there. A bone with no weights of its own still moves the mesh if its children carry weight, or if a Mirror modifier builds the other half — so those bones are named separately, not accused. A character made of a body, gloves and eyes on one rig is read as one character: a bone is only called empty when it carries no weight on any of them. A mesh bound by envelopes is not called unbound, because weights are not what moves it.

And where the measurement itself cannot work — the armature is in Rest Position, or it is a control rig whose bones are driven by constraints — it says so and judges nothing. The zero it prints always sits next to a number that is not zero.

It does not write to your weights

Not one of these is fixed. There is a shelf of add-ons that write weights, including our own — and a tool that fixes cannot be the witness that the fix worked. BindProof reads the bind you already have, whoever made it: Blender's binder, Voxel Heat, an imported FBX, your studio's own pipeline.

A fingerprint that includes the weights themselves and the whole pose is taken before and after every check and compared, and the panel prints the result each time — not only when something went wrong.

Tested

108 checks on each of Blender 4.2.9, 5.1 and 5.2, zero problems, plus ten deliberate sabotages of the product itself to prove the tests can fail — all ten turn the suite red, including ones that treat a zero weight as a bind, that stop telling deforming bones from ordinary vertex groups, and that look for the armature through parenting instead of through the modifier.

Silence is tested as hard as noise: on a fully bound mesh it reports nothing.

Requires

Blender 4.2 or newer. No internet connection, no external libraries, nothing written to your file system.

---

And now it fixes what the engine would drop

Unity, Unreal and Godot read four influences per vertex and silently discard the rest, so a character can deform one way in Blender and another way in the game. BindProof already measured how far apart those two are. Now one button keeps the four heaviest influences, makes the weights add up to one, and measures again.

The proof is not "the warning went away". It is that the prediction and the result agree: BindProof said a vertex would sit 893.9 mm out of place in the engine, the repair ran, and the skin in Blender moved by 893.9 mm — a difference of 0.0001 mm. What you see is now what ships.

Unbound vertices and empty bones are deliberately left alone. Which bone should own a vertex is your decision; guessing "the nearest one" is how an ear ends up sewn to a shoulder. BindProof names them and shows you where they are.

---

Whole deliveries, not one file at a time

• Check a folder. Point it at a delivery and every .blend inside is checked, each in its own background Blender. The scene you have open is not touched — not opened over, not saved, not changed. A file that hangs or crashes takes only its own process down; the other thirty-nine still get checked.
• A report you can send. One self-contained page: verdict first, one sentence a lead can read, then the numbers, then every finding with its address. No internet, no fonts to load, no dependencies — it opens in six months on a machine with no network, which is exactly when someone asks what you delivered.
• Runs without a person. A single command from your pipeline writes the report and returns a code: clean, something blocking, or could not be checked. Those are three different codes on purpose — a check that did not run has not passed, and a gate that treats them the same is worse than no gate at all.
• Remembers last time. Save today's numbers as a baseline and a later run tells you what moved, in which direction, and which findings are new. A measurement that went up is not automatically bad: each product says which way is worse for its own numbers.
• One verdict for the whole delivery. If you own more than one of these tools, one run puts every check you have over every file and answers the only question that matters before you send it: can this go out? Each file is opened once and all the checks see the same state of it.