We use FrenetSerretSystem to calculate its tangent vector and normal vector at c[t0].

c[t_] = {t^2, t};
tangent[t_] = FrenetSerretSystem[c[t], t][[2, 1]];
normal[t_] = FrenetSerretSystem[c[t], t][[2, 2]];
s[t_] = ArcLength[c[τ], {τ, 0, t}];
t0 = .5;
Show[ParametricPlot[c[t], {t, 0, 1}], 
 Graphics[{Arrow[{c[t], c[t] + normal[t]}], 
    Arrow[{c[t], c[t] + tangent[t]}]}] /. t -> t0, PlotRange -> All]

And then we calculate the ArcLength of the parametric curve {t^2,t} from 0 to t0 then we get the point {0,s[t0]} in y-axis.

After that we translate {0, s[t0]} - c[t0] and rotate tangent[t0] to the y-axis direction {0,1} so that the new curve tangent to the y-axis at {0,s[t0]}.

c[t_] = {t^2, t};
s[t_] = ArcLength[c[τ], {τ, 0, t}];
r[t_, t0_] := 
  RotationTransform[{c'[t0], {0, 1}}, {0, s[t0]}][
   c[t] + {0, s[t0]} - c[t0]];
Manipulate[
 ParametricPlot[r[t, t0], {t, 0, 1}, AspectRatio -> Automatic, 
  PlotRange -> {{0, 1}, {0, 2}}], {t0, 0, 1}]

New Edition

We general the idea from above to deal with two parametric curves. Here we use NDSolve to consider the re-parametric equation, and thanks @Daniel Huber provide FunctionInterpolation to increasing the speed.

r1[t_] = {1.5 + 1.5 Cos[π - t], Sin[π - t]}; t1 = 
 FunctionInterpolation[
   InverseFunction[
     NDSolve[{s1'[t] == Norm[r1'[t]], s1[0] == 0}, 
       s1, {t, 0, 100}][[1, 1, 2]]][x], {x, 0, 50}] // Quiet;
r2[t_] = {t^2, 3 t};
t2 = FunctionInterpolation[
    InverseFunction[
      NDSolve[{s2'[t] == Norm[r2'[t]], s2[0] == 0}, 
        s2, {t, 0, 100}][[1, 1, 2]]][x], {x, 0, 50}] // Quiet;
Animate[Show[ParametricPlot[{r1[t1[s]], r2[t2[s]]}, {s, 0, 20}], 
  Graphics[Arrow[{r1[t1[s]], r1[t1[s]] + D[r1[t1[s]], s]} /. 
     s -> s0]], 
  Graphics[Arrow[{r2[t2[s]], r2[t2[s]] + D[r2[t2[s]], s]} /. 
     s -> s0]]], {s0, 0, 20}, DefaultDuration -> 10]

r1[t_] = {1.5 + 1.5 Cos[π - t], Sin[π - t]}; t1 = 
 FunctionInterpolation[
   InverseFunction[
     NDSolve[{s1'[t] == Norm[r1'[t]], s1[0] == 0}, 
       s1, {t, 0, 100}][[1, 1, 2]]][x], {x, 0, 50}] // Quiet;
r2[t_] = {t^2, 3 t};
t2 = FunctionInterpolation[
    InverseFunction[
      NDSolve[{s2'[t] == Norm[r2'[t]], s2[0] == 0}, 
        s2, {t, 0, 100}][[1, 1, 2]]][x], {x, 0, 50}] // Quiet;
trans[ss_, ss0_] := 
  RotationTransform[{D[r1[t1[s]], s], D[r2[t2[s]], s]} /. s -> ss0, 
    r2[t2[ss0]]][r1[t1[ss]] + r2[t2[ss0]] - r1[t1[ss0]]];
curves = ParametricPlot[{r1[t1[s]], r2[t2[s]]}, {s, 0, 20}];
Animate[Show[curves, 
  ParametricPlot[trans[ss, ss0], {ss, 0, 3 π}]], {ss0, 0, 20}, 
 DefaultDuration -> 10]