@Rikkes wrote:
.... Is it possible to make it work on arcs to? ....
One of the questions that arises: What should be the criterion for the cut-off points on the initial Arcs relative to their intersection? [This is assuming, because of your description of the need, that they do intersect, or their extensions do.] As an illustration, using Arcs of significantly different radius to illustrate how it matters:

Calling the point where the white Arcs meet the "peak," and with the green and yellow Arcs of equal* radius:
The green Arc's ends are at equal chord lengths along the white Arcs [straight-line distances] from the peak. The yellow Arc's ends are at equal distances along the white Arcs from the peak. [A similar difference would apply when done between an Arc and a Line.]
Is one or another of those, or maybe some other criterion, what you would want used to define this kind of bubble fillet, for the right kind of end result?
* They're not actually precisely the same radius, but very close, for illustration. Alternatively, I could have made them the same, and slightly fudged the "equality" of the chord or along-the-Arcs distances. This is one of the complications, depending on what criterion you choose -- it may not be possible to calculate it with exact precision, but could require actually drawing a temporary Circle of desired radius passing through the peak, testing the result at its intersections with the initial Arcs for equality of the criterion distances, fine-tune-Rotating the Circle about the peak in a direction to get them closer to equal and testing again, repeatedly until the variance from precise equality falls within some small tolerance.
Kent Cooper, AIA