# Filtering private API

**URL:** <https://discourse.julialang.org/t/filtering-private-api/126349>\
**Category:** General Usage\
**Tags:** question, documentation, documenter\
**Created:** [February 26, 2025, 4:28pm UTC](https://discourse.julialang.org/t/filtering-private-api/126349 "2025-02-26T16:28:46Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![s\_amap](https://avatars.discourse-cdn.com/v4/letter/s/b19c9b/32.png) [@s\_amap](https://discourse.julialang.org/u/s_amap)\
**Post date:** [February 26, 2025, 4:28pm UTC](https://discourse.julialang.org/t/filtering-private-api/126349/1 "2025-02-26T16:28:47Z")

</div>

A bit of a naive question regarding documenter indexing :

I have a package with a public API, and some internal developer functions that are mostly of no interest to users, with a few potential exceptions for power users.

I’d like to create a page detailing some of the internals for those exceptions. But there doesn’t seem to be a straightforward way of tagging functions with some tag that could then be used with `@index` in some way or another as far as I can see ?

The way I’ve thought of going about this is to create a dummy module, export the relevant functions and use that module as a filter.

Is this the standard way of doing this or am I missing some more straightforward way of accomplishing the same goal ?

---

<div class="post-metadata">

**Author:** ![s\_amap](https://avatars.discourse-cdn.com/v4/letter/s/b19c9b/32.png) [@s\_amap](https://discourse.julialang.org/u/s_amap)\
**Post date:** [February 27, 2025, 9:54am UTC](https://discourse.julialang.org/t/filtering-private-api/126349/2 "2025-02-27T09:54:55Z")

</div>

I also have an even more awkward use case :

There is a method exported with two different ‘main’ user entry points which are documented, but it has other overloads, one of which might be occasionally useful to power users.

I’d like to document that function without interfering with the public API, and without hard-writing it into the documentation, and I don’t think the module approach works there, so I’m throwing that bonus question in the mix.