camberay

Toe · camber · caster · SAI · thrust — every wheel, one solve

Align your own car. Properly.

No strings, no spirit levels, no €100 shop visit.2 Coded targets on your garage floor, your phone, and every angle on the screen — live, while you turn the wrench. Print the targets on any printer, 3-D print them, or buy them made: the software measures your layout, so the result is the same.4

One email when you can order. One-click unsubscribe. Privacy.

Spec sheetrev. 2026-10
Angles
toe · camber · caster · SAI · thrust
Per wheel
all four, plus both axles
Toe
up to 0.05° 1
Camber
up to 0.05° 1
Caster · SAI
≤ 0.5° 1
Resolution
0.01°
Reference
your floor, measured — not gravity
Floor
flat; does not need to be level
Setup · capture
~10 min · ~20–30 min
Results
on the phone, live per wheel + A4 report
Camera
an Android phone, calibrated in the app · iPhone app soon
Targets
any printer + cardboard · 3-D print · made
Licence
made kit €549, the app for life · print it yourself: €149 for life, €59 a year, €29 for 30 days · VAT incl.
Use
personal: unlimited measurements of up to 5 cars
The measurement setup, drawn to scale A perspective line drawing of a car on a faint floor grid, surrounded by round coded targets on squares of paper and A4 tiles raised on legs of varying height, with a coded target plate on each wheel. Callouts show the camber and toe measured at the front-right and rear-right wheels, and the car's thrust line. FRcamber -0.40° · toe -0.08° RRcamber -0.60° · toe -0.03° THRUST

The setup, drawn to scale and projected through a real pinhole camera — 2.6 m wheelbase, 1.5 m track, tiles on legs, anchors clear of the jack points.

No garage tricks

Nothing to square up, level or eyeball.

Strings and jack stands. Tape measures across the tyres. A spirit level on the rim. A perfectly level floor. The software measures the floor you have and every wheel against it — including the two angles the string box can never give you, caster and SAI.

No shop bill

Measure as often as you like, in your own garage.

A four-wheel alignment costs €70–120 a visit2 and nobody gets the numbers right first time. Set it, drive it, measure again the same evening — the loop that used to cost three appointments now costs 30 to 40 minutes.

Targets your way

Office printer and cardboard. 3-D printer. Or buy them made.

Same software, same accuracy.4 The targets only have to be flat and rigid, not precise: the solve measures your actual printed layout and calibrates itself to it. Pick the version that suits your evening, not your budget.

What you get back

One page. Every wheel. And a number it will refuse to print.

Everything you see live on the phone also lands on a finished A4 sheet per measurement, so you can put two side by side and see what your adjustment did. Here is the layout, with a car that has a real problem in it.3

Sample alignment report sheet A one-page report: a top view of a car with each wheel coloured by status, a measurements table giving camber and toe per wheel against the spec band, total toe front and rear, thrust angle, and front caster and SAI. The front wheels are out of spec. WHEEL ALIGNMENT REPORT hot hatch, 18″ · 2026-07-27 · own driveway SAMPLE FRONT FL camber -1.68° toe +0.31° RL camber -1.22° toe +0.05° FR camber -1.55° toe +0.28° RR camber -1.31° toe +0.09° MEASUREMENTS WHEELCAMBERTOE STATUSEVIDENCE FL -1.68° +0.31° OUT OF SPEC 41 pts FR -1.55° +0.28° OUT OF SPEC 38 pts RL -1.22° +0.05° PASS 44 pts RR -1.31° +0.09° PASS 36 pts TOTAL TOE FRONT +0.59° TOTAL TOE REAR +0.14° THRUST ANGLE -0.02° STEERING GEOMETRY (FRONT) FL · caster +4.12° · SAI +11.8° FR · caster +4.35° · SAI +11.6° Angles are RELATIVE TO THE FLOOR the car sits on — measured from the floor markers, not from gravity. Spec band (default passenger car): camber -1.5…+0.5°, toe -0.10…+0.20°. Camberay · CV wheel-alignment kit
Illustration of the report layout, not a measurement of a real vehicle.

It reads the car, not the wheel

Total toe front +0.59° with a straight rear axle. A toe gauge would have told you the front pair were toed in. It would not have told you the rear is fine — which is how you know the problem is in front of you.

Thrust is in the box, not an upsell

A thrust of −0.02° means the rear axle points where the car goes. When it doesn't, the steering wheel sits crooked on a straight road and every front number you set is measured off a lie. Most tools in this bracket cannot see it at all.

And when it isn't sure, it says so

A wheel that fails its own consistency checks is printed like this — no number, a plain reason, and what to re-shoot:

RR re-measure plate moved between frames — re-shoot RR from one step back no unverified value is reported

Refusing beats guessing. A tool that always produces a confident number is not more accurate than one that sometimes doesn't — it is just quieter about being wrong.

On the phone, live

One of you watches the screen. The other turns the wrench.

No waiting for a sheet. Point the camera at one wheel and its camber and toe update live on the phone while somebody adjusts the tie-rod, so you stop when the number says stop, not after the third drive round the block.

Live per wheel

Camber and toe for the wheel in view, refreshed continuously against the floor map you just built. Set your target and watch the bar walk onto it.

Whole car at a glance

The other three wheels, total toe and thrust stay on screen, so an adjustment on one side is seen for what it does to the rest of the car.

Freeze it, print it

When it reads right, freeze the state and the A4 report is generated from that moment — the same page you can put next to last month's.

LIVE · FRONT LEFT 12 fps CAMBER −1.42° target −1.30 TOE +0.06° in spec OTHER WHEELS FR camber −1.31° toe +0.05° RL camber −1.22° toe +0.05° RR camber −1.31° toe +0.09° TOTAL TOE FRONT +0.11° THRUST −0.02° FREEZE & PRINT REPORT

How it goes

Ten minutes to place. 20–30 to shoot. The rest is software.

01 · PLACE

Targets on the floor

Stand the tiles around the car, scatter the anchors between them, clip one plate to each wheel, and hold the steering wheel straight. Nothing has to be square, level or measured.

02 · WALK

Two slow laps with the phone

Walk around the car filming the floor. The app shows live coverage and tells you where the map is still thin.

03 · SPIN

Jack, spin, set it down

Jack each wheel in turn and spin it while the phone films the plate. Then all four down, a short roll back and forth to let the tyres go, and photos of each plate as the car stands.

04 · READ

On the phone, then on paper

Every wheel's toe and camber, both total toes and the thrust angle — plus front caster and SAI with the optional steering sweep — on the screen, and on an A4 sheet when you freeze it. Adjust, and watch it move.

Step by step, with placement rules and what the app asks of you: the field guide.

Why a tenth of a degree

Toe you can't feel is sanding your tyres flat.

A wheel that doesn't point where the car is going gets dragged sideways for every metre you drive. The arithmetic is exact: sideways drag equals the tangent of the toe angle, times the distance travelled. Half a degree is visible in a tyre after a season. A tenth is what separates a car that turns in evenly from one that doesn't.

How toe angle turns into sideways tyre scrub A top view of one wheel pointing slightly inward of the car's direction of travel, and a bar chart of the resulting sideways drag per kilometre: 0.10 degrees gives 1.7 metres, 0.25 degrees gives 4.4 metres, 0.50 degrees gives 8.7 metres. ONE FRONT WHEEL, SEEN FROM ABOVE SCRUB where the wheel points where the car goes angle exaggerated ~50× to be visible SIDEWAYS DRAG PER KILOMETRE DRIVEN — ONE TYRE 0.10° 0.20° total toe 1.7 m 0.25° 0.50° total toe 4.4 m 0.50° 1.00° total toe 8.7 m drag = tan(toe) × distance 0.25° per wheel, over a 15 000 km year: 65 km of pure scrubbing, per tyre.
This is why the interesting resolution is a tenth of a degree, not half a degree. Half a degree is visible in a tyre after a season. A tenth of a degree is what separates a car that turns in evenly from one that doesn't — and no gauge that reads to 0.5° can tell you which one you have.

The real problem

Alignment is a loop. A shop visit is a straight line.

Nobody gets it right in one pass. You set a number, you drive it, you change your mind. Every turn of that loop currently costs an appointment, a drive and a morning — so most people run it once and live with the result for four years. A camber bolt takes four minutes and a tie-rod takes ten; knowing what the adjustment did is the part you pay someone else for.

A shop visit compared to measuring on your own floor On the left, a one-way chain: book a slot, drive there, one measurement, drive home. On the right, a closed ring: measure, adjust, confirm, repeating on your own floor. THE SHOP Book a slot Drive there One measurement Drive home €70–120. Changed your mind? Start at the top. YOUR FLOOR MEASURE ADJUST CONFIRM on your own garage floor Adjusting was never the hard part. Closing the loop was.

Against the shelf

Everything else measures a wheel. This measures a car.

The DIY shelf is full of good, honest tools that each read one or two angles off one reference. No single instrument gives you the whole geometry at once, so you end up owning three of them and reconciling by hand.

Categories, not brands — capability as each class of tool is normally sold.
Tool Toe, per wheel Camber Caster & SAI Thrust Reference it trusts
Tape-and-plate toe gauges / laser toe bars total only — — — the rim you clamped to
Digital camber-caster gauge — yes with a steering sweep — gravity — so the floor must be level
String box yes — — yes the frame you squared up by hand
Phone app on handset sensors yes yes yes indirect your handset's IMU and magnetometer
Camberay yes yes yes yes your own floor, measured on the spot

Nothing to level

Angles are reported against the floor the car stands on, whatever its slope. A drainage fall is fine. Steps and crowns are not.

No magnetometer to fool

Sensor-based phone tools inherit whatever IMU and compass the handset shipped with, and a steel floor or a brake rotor bends a magnetic heading. A camera measures geometry it can see.

It knows when it's wrong

Every wheel is cross-checked before a number is printed. That is why the report can say re-measure instead of quietly giving you a confident wrong answer.

What the software does for you

You place. You shoot. It handles everything that used to need care.

Every step that makes garage alignment fiddly — squaring things up, levelling, mounting a gauge dead straight — is done in software from the photographs. You only have to be roughly right.

01 · PLACEMENT

Rough is fine

Targets go down wherever there is floor. No grid, no measuring tape, nothing has to be parallel to anything.

02 · FLOOR

Slope doesn't matter

Angles are read against the floor the car actually stands on. A drainage fall or a sloping drive gives the same numbers as a level bay.

03 · MOUNTING

Crooked plate, right answer

Clip the wheel plate on by hand. However it sits — tilted, off-centre, on a bent rim — the software compensates and reports the wheel, not the plate.

04 · CHECKING

It grades its own work

Every wheel carries a confidence check. Pass, and you get the number. Fail, and you get re-measure with what to re-shoot — never a guess.

Worth being blunt about

Every clamp-on tool measures the rim and its adapter, and you inherit whatever runout and mounting error came with them. This measures the wheel as it actually turns, so a plate mounted slightly crooked stops being your problem.

What you place

Targets on the floor, a plate on each wheel.

Nothing in the kit needs to be precise — no machined bar to keep flat and unbent in a garage, no exact placement. Put the targets down roughly where the app shows you; the software takes it from there.

Floor tiles

×20

A4-sized target faces on short legs, stood around the car. Cardboard, 3-D print or made — your choice.

Anchors

×112

Single coded targets lying flat between the tiles. Scatter them; the app tells you where more are needed.

Wheel plates

×4

One target plate per wheel, held on by a clip-on holder. It does not need to sit square or centred — the software compensates for how it was mounted.

Plate holders & turn plates

×4 · ×2

Holders that clip a plate to any wheel, and two turn plates for the front wheels to sit on for the steering sweep that gives caster and SAI.

Your phone

×1

An Android phone; the iPhone app is soon. Its camera is calibrated once against a printed sheet — guided in the app, checkable afterwards. This is the one hard requirement.

It packs into a document box and goes down in about ten minutes. Nothing is precious — a scuffed anchor still reads, and a target you tread on is replaced for the cost of a print. Compare that with owning a straightedge you are afraid to drop.

Three ways to make the targets

One software. Targets any way you like.

The targets carry no precision of their own — the software measures the layout you actually made and calibrates itself to it. So make them the way that suits you: the result is the same.4

Cheapest

Office printer + cardboard

  • Print the PDFs on any A4 printer
  • Glue to cardboard or foam board, cut, done
  • Tile legs from offcuts, boxes or books
  • An evening with scissors and glue

Same accuracy as the other two.4 Flat and rigid is all it asks.

Most durable

3-D print it

  • STL files for the plates, legs and mounts
  • Print in any colour — the detector doesn't care
  • Survives being kicked, stacked and driven over
  • Filament and an afternoon of printer time

For people who own a printer and like a project.

Nothing to make

Buy it made

  • Tiles, anchors and plates printed, cut and ready
  • Packs into a document box
  • Place it and shoot
  • The app for life comes with it
  • Made to order in small batches

€549, EU shipping included. On a reservation list now: no payment is taken.

Made kit

€549, EU shipping included, with the app for life. It is on a reservation list now: no payment is taken, and the list hears first when it can be ordered.

Print it yourself

Without the made kit, the app and the files to print or 3-D print the targets come three ways: €149 for life; €59 a year, renewed each year until you cancel (cancel any time, and it ends with the year you paid for); or a €29 pass for 30 days — one payment, and it does not renew.

Personal use

Every plan, the made kit included, is for personal use: unlimited measurements of up to 5 cars. A workshop, a club or anyone measuring more cars needs the commercial licence.

Private

Your photos stay on your phone. A photo leaves it only if you join the photo beta and agree to it.

Connection

Taking the photos needs no connection. The result needs one.

Verifiable

Every result shows its own confidence per wheel, and the app will not print a number it could not cross-check. You never have to take a reading on faith.

Prices in euros, VAT included. No orders are taken now: join the waitlist.

To be plain about it

The printable files are not a free download; they come with what you buy. And a claim we do not make: "open". The software is ours. What is yours is every number it prints, with the confidence it printed it at — which is the part worth trusting.

Questions

The obvious ones.

How accurate is it?

Toe and camber up to 0.05°, caster and SAI to within 0.5°, reported at 0.01° resolution.1 A wheel is printed only when its own consistency checks pass; otherwise the sheet says re-measure and tells you what to re-shoot. How the figures are tested: how we test accuracy.

Worth knowing while you shop: many figures printed on this category's boxes are resolution — the smallest change the display can show — which is a much easier number to earn than accuracy, and not the same promise.

How is this different from a phone gauge or a string kit?

A gauge measures one wheel against gravity, and strings measure against a frame you built by hand. This measures every wheel against your own floor, all four in one go, and cross-checks each result before it prints it — so it can tell you when a reading is not trustworthy instead of leaving you to find out on the road.

Does my garage floor need to be level?

Flat matters; level doesn't. The map is built from the floor targets themselves, so a drainage fall or a gentle slope is fine — the angles are reported against that measured plane. A floor with steps in it, or a big crown, will bias the result. Park somewhere reasonably flat with about a metre clear around the car.

Do I need a 3-D printer?

No. Print the target PDFs on an ordinary A4 printer, glue them to cardboard or foam board and cut them out; prop the tiles on anything rigid. The software measures the layout you made rather than assuming a nominal one, so paper-and-cardboard targets give the same result as printed or made ones.4 The 3-D printed and made versions are simply tougher and quicker to set up.

Is there a subscription?

Only if you choose one. The made kit (€549, EU shipping included) comes with the app for life. Without it, the app and the printable files cost €149 for life; €59 a year, which renews each year until you cancel — cancel any time, and it ends with the year you paid for; or €29 for a 30-day pass, one payment that does not renew. Prices in euros, VAT included.

Every plan is for personal use: unlimited measurements of up to 5 cars. For more cars, there is the commercial licence.

Do I need a lift, or stands?

A trolley jack. Each wheel comes off the ground in turn to be spun, then goes back down. With all four on the floor, a short roll lets the tyres settle, and the angles are read from photos of the car as it stands — the state you actually want to measure.

What happens if it can't read a wheel?

It tells you which wheel, why, and where to stand for the retry — then reports re-measure with no number rather than a confident guess. The causes are dull and fixable: a plate that moved, a dark corner of the garage, or a wheel with too little floor visible behind it.

Do my photos get uploaded somewhere?

No. Your photos stay on your phone. A photo leaves it only if you join the photo beta and agree to it. What the app sends us, and why, is in the privacy notice.

Is it open source?

No, and we do not imply otherwise. The software and the printable files come with what you buy. What you can always check is the result: every wheel on every report carries the confidence it was measured at, and a wheel that fails its check is never given a number.

Will it tell me what my car's numbers should be?

It measures your car and checks it against a general passenger-car band. You bring the factory spec, which for most cars is a search away and for a track car is a decision you were going to make yourself anyway.

Waitlist

One email when you can order.

Nothing in between. The list also gets first pick of the made kits.

Notes

  1. Toe and camber: up to 0.05° is from preliminary field tests of one car with one phone, measured on two days. On the second day, read again from parts of its own photographs, each wheel's toe moved by 0.01–0.05° and its camber by 0.02–0.06° (one standard deviation); two separate reads of the same photographs, each with its own target detection and floor map, agreed within 0.03°. The total toe of each axle differed by 0.03° or less between the two days. These tests show how a read repeats on that car; no alignment machine measured it. Caster and SAI: the figure is from simulation on rendered imagery. These figures are not a guarantee of the result on any particular car or floor. The method of the full study: how we test accuracy. Every angle is reported to 0.01° resolution; resolution is not accuracy.
  2. Typical quoted price of a four-wheel alignment at independent and franchise workshops in the EU and US, 2026. Your local price will differ.
  3. The report and the phone screen shown are illustrations of the layout with representative values, not measurements of a real vehicle.
  4. The solver does not rely on the printed geometry being exact: it measures the position of every target from the photographs and calibrates each tile's true shape, so print scale, paper stretch and how you cut the cardboard drop out. What the targets must be is flat and rigid for the duration of the measurement. The same software reads paper-on-cardboard, 3-D printed and made targets.