The \emph{general} construction of the center of the circle that is tangent to two given circles $\Gamma_1$ and $\Gamma_2$ (with centers $M_1$ and $M_2$ and radii $r_1$ and $r_2$ respectively) and a line $\ell$ where $\Gamma_1$ and $\Gamma_2$ are on the same side of $\ell$ that I found is pretty complicated, but bear with me:
\begin{enumerate}
\item Construct the line $k_1$ as the line on the other side of $\Gamma_1$ with $d(k_1, \ell) = r_1$ and $k_1 \parallel \ell$, and construct $k_2$ analogously.
\item Let $p_1$ be the parabola with focus $M_1$ and directrix $k_1$, and define $p_2$ analogously.
\item The intersection point of $p_1$ and $p_2$ is the desired center of the circle that is tangent to $\Gamma_1$, $\Gamma_2$ and $\ell$.
\end{enumerate}
This works since all points $P$ on $p_1$ suffice $d(P,\Gamma_1) = d(P,\ell)$ and similar. \\ \quad \\
In your case, however, $\Gamma_1$ and $\Gamma_2$ have the same radius $r$ and are of equal distance to $\ell$. In that case, there's nothing holding you back doing some Pythagoras and get the radius $R$ of the wanted circle, then getting the intersection of the circles with centers $M_1$ and $M_2$ with radius $r + R$. (Now that I think about it, the same might be possible for the general case as well... Let me think about that.)
https://cdn.discordapp.com/attachments/757583324632973403/937862098753568798/unknown.png


that's an equal sign?




