Incircle and Circumcircle of a Triangle Calculator
Every triangle has two remarkable circles: the incircle, tangent to all three sides from the inside, and the circumcircle, which passes through all three vertices. Enter the coordinates of the three vertices A, B, C: the page works out the center and radius of both circles on a single diagram.
Coordinates of the three vertices
An asterisk after a term points to its entry in the glossary below.
Two circles, two jobs: incircle versus circumcircle
Two circles, one triangle, two jobs. The circumcircle* runs through all three vertices. Its circumcenter* O is equally far from A, B and C; that common distance is the circumradius* R. The triangle sits inside; the circle wraps it.
The incircle* touches all three sides, each at one point. Its incenter* I is equally far from the three supporting lines; that common distance is the inradius* r. The circle sits inside; the triangle wraps it.
O and I coincide only in the equilateral case. Anywhere else they drift apart. On an obtuse triangle O even leaves the figure — the circumcircle spills well outside. I never does; the incenter stays inside. This page finds both from vertex coordinates, not from side lengths alone. The sides come afterwards, and so does the area, via the shoelace formula. The unit dropdown is a label: coordinates 0, 8, 3 stay those numbers; the radii come back in the named unit.
Why the perpendicular bisectors meet
The perpendicular bisector of a segment is the set of points equally far from its two ends. That is a definition, not a theorem. It is enough.
- On the perpendicular bisector of AB, every point M satisfies MA = MB.
- On that of AC, every point N satisfies NA = NC.
- The two lines meet (they would be parallel only if B, A, C were collinear, which is not a triangle). Call the intersection O.
- Then OA = OB and OA = OC, so OB = OC. O also lies on the perpendicular bisector of BC.
All three lines share a point. That point, equidistant from the vertices, is the centre of the only circle through them. The third bisector need not be drawn to find it: two suffice, the third follows. The same construction, run on three stakes in the field, is how a surveyor recovers a point equally far from three stations — often without naming the bisectors at all.
Why the angle bisectors meet
Equidistance now talks about sides, not vertices. A point on the interior bisector of an angle is equally far from the two sides of that angle. That is the bisector property used with compass and straight-edge.
- On the bisector of angle A, every point is equally far from line AB and line AC.
- On that of angle B, every point is equally far from BA and BC.
- Their intersection I is therefore equally far from all three sides. A circle centred at I touches all three.
- That common distance also puts I on the bisector of angle C.
Again, two bisectors fix I, the third confirms. The inradius r is exactly that common distance. It can also be written r = Area / s without ever drawing I — the shortcut on the sister triangle area and perimeter calculator, which starts from the three lengths. Coordinates are needed here because the centres have to be placed, not only sized.
Area from coordinates: the shoelace formula
Three points A(x₁, y₁), B(x₂, y₂), C(x₃, y₃). The area, with no sides required:
Hence the name shoelace*. The cross products xi yi+1 minus xi+1 yi are “laced”, the loop is closed back to the first vertex, and the absolute value is taken. It is Gauss's area formula for a general polygon, cut down to three vertices.
Geometrically each cross product xi yi+1 − xi+1 yi is twice the signed area of the small triangle formed by the origin and one side. Going round, the parts lying outside triangle ABC cancel in pairs, leaving twice the area, with a sign that records clockwise versus anticlockwise order. The absolute value throws the sign away. This page needs that: the vertices arrive as coordinates, not as lengths. The distance between two points then supplies the three sides, once the vertices are known.
Three collinear points give area zero. Shoelace spots it, and the calculator stops — no incircle of radius zero, no infinite R.
Right-angled and equilateral special cases
In a right triangle the right angle subtends a semicircle. Thales' theorem — the converse everyone actually uses — says the hypotenuse is a diameter of the circumcircle. So O is the midpoint of the hypotenuse, and R = hypotenuse / 2. Quick check on a 3-4-5: R = 5/2 = 2.5, and abc/(4A) = 60/24 = 2.5. Matches.
The equilateral stacks every centre on the same spot: O, I, orthocentre, centroid. R = 2r, with r = a√3 / 6 and R = a√3 / 3. It is also a regular 3-gon: the polygon's apothem is r; its circumradius is R.
In between, a non-equilateral isosceles keeps O and I on the median from the apex, but they no longer sit on top of each other. The gap OI is Euler's formula, below. A full set of angles from the three sides, if that is the next question, is the triangle solver.
Euler's identity OI² = R(R − 2r)
Euler left a short identity tying the distance of the two centres to the two radii:
OI² = R(R − 2r) = R² − 2Rr
Same identity, two writings. Immediate corollary: R ≥ 2r, always. Equality, OI = 0, happens only when R = 2r, that is, in the equilateral. Anywhere else R overshoots 2r, and O drifts from I.
On a 3-4-5, r = 1, R = 2.5, so OI² = 2.5 × (2.5 − 2) = 1.25, OI ≈ 1.118. The centres are not far; the triangle is not needle-thin. On a nearly flat triangle r collapses, R blows up, and so does OI. Euler's formula is a thermometer for how close the figure sits to equilateral — still a number, not a judgement.
Inscribed angles and Thales' theorem
An inscribed angle is half the central angle that stands on the same arc. So every inscribed angle looking at the same arc is equal. In a triangle, the angle at A stands on arc BC, hence A = ½ of the central angle BOC.
Famous special case: an inscribed angle standing on a diameter is a right angle. Thales again. That is why O is the midpoint of the hypotenuse. The converse is a right-angle test that never measures the angle: if O lands at the midpoint of the longest side, the opposite angle is 90°.
The same theorem sorts cyclic quadrilaterals (opposite angles supplementary). Not this page's brief, but a triangle's circumcircle is the first circle of a family. Three points, always a circle; four points, only sometimes. The circle itself is that circumcircle with the triangle stripped away: circumference 2πR, area πR².
The two working identities r = A/s and R = abc/(4A)
Once the area is known, the radii no longer need the coordinates of I and O.
The first is visible on the sketch: the three small triangles with vertex I have areas ½ar, ½br, ½cr, summing to r·s. The second comes from A = abc / (4R), itself from A = ½bc sin A and a / sin A = 2R (the sine rule).
This page gets the area by shoelace, then applies those two lines. They work with no coordinates at all, as soon as the three sides are known — that is the bridge to the Heron calculator. The centres, though, stay an analytic-geometry problem: I is the side-weighted average of the vertices, O is the cut of two perpendicular bisectors.
Where the circumcenter sits: acute, right, obtuse
O's seat depends on the angles, not on the sides taken one at a time.
- Three acute angles. O lies inside the triangle, as I does.
- One right angle. O is the midpoint of the hypotenuse, so on the boundary.
- One obtuse angle. O steps outside the triangle, beyond the longest side, on the side away from the obtuse angle. The circumcircle swallows the triangle and spills well behind that longest side.
I never does that. The incenter stays interior, always. On the diagram, if O appears outside the triangle, look for the obtuse angle — it is opposite the longest side. Shoelace does not care: the area stays positive, r and R as well. Only the picture changes shape.
Checking a 3-4-5 without a coordinate frame
No axes required. Sides 3, 4, 5, s = 6, area 6. Then r = 6/6 = 1, R = 60/24 = 2.5. Clean numbers. Park B at (0, 0), C at (3, 0), A at (0, 4) for the picture: I lands at (1, 1) — distance 1 from both axes, as it must. O is the midpoint of the hypotenuse, (1.5, 2).
OI² = (1.5 − 1)² + (2 − 1)² = 0.25 + 1 = 1.25, and R(R − 2r) = 2.5 × 0.5 = 1.25. Euler, checked by hand, no library. That is the sanity-check triangle for this page. If a result ever comes out elsewhere on a 3-4-5, it is the code, not the geometry.
The same three vertices can be entered on the form as a check: A(0, 4), B(0, 0), C(3, 0). Inradius 1, circumradius 2.5, area 6. Anything else, and a coordinate has slipped. The rest of plane geometry on this site starts from sides rather than vertices; this page is the one that needs the frame.
Glossary
- Incircle
- The unique circle tangent to all three sides of a triangle, lying entirely inside it. Its radius is the inradius r.
- Incenter (I)
- The centre of the incircle: the point where the three interior angle bisectors meet, equally far from all three sides. (US spelling; UK: incentre.)
- Inradius (r)
- The radius of the incircle. Equal to the area divided by the semi-perimeter: r = A / s.
- Circumcircle
- The unique circle that passes through all three vertices of a triangle. The triangle is inscribed in it.
- Circumcenter (O)
- The centre of the circumcircle: the point where the three perpendicular bisectors meet, equally far from all three vertices. (US spelling; UK: circumcentre.)
- Circumradius (R)
- The radius of the circumcircle. Equal to abc / (4A). In a right triangle it is half the hypotenuse.
- Shoelace formula
- Gauss's area formula for a polygon given by vertex coordinates. For a triangle it reduces to A = ½ |x₁(y₂ − y₃) + x₂(y₃ − y₁) + x₃(y₁ − y₂)|.