I am looking for a rigorous, textbook-style proof of the Laplace approximation of general exponential integrals of the form $$ \int e^{\lambda h(\lambda,t)}g(\lambda,t)\,dt \quad\text{as}\quad \lambda\to\infty. \tag{*} $$
Every treatment I have found assumes that $h(t)$ does not depend on $\lambda$. Frustratingly, Miller's "Applied Asymptotic Analysis" introduces the subject using the general integral $(*)$, but only proves results for $h(t)$.
(There are many more resources and lecture notes online, e.g. these notes, but they only treat $h(t)$ as well.)