In current (mono-temporal) Anchor Modeling we push the "recording time" down to metadata. The tool will not generate a metadata model for you, it will have to be modeled as well. Ideally, your metadata model is also an anchor model. The connection between your data model and your metadata model is through the metadata identifier found on every row in every table.
For example, say you have a source system 'A' that on the 20th of August 2001 delivered the information: "The price of product 42 is 10 euro in effect through all of 2001, and after that 12 euro". Your data would be <#42, 10, [2001-01-01, 2001-12-31]>, <#42, 10, [2002-01-01, ~]> and your metadata is <'A', 2001-08-20> before storing it in an anchor model. Now, when inserted into the data model and metadata model, you get:
Price attribute
(PR_ID*, PR_PRI_Product_Price, PR_PRI_ChangedAt*, PR_PRI_Metadata)
<#42, 10, 2001-01-01, #555>
<#42, 12, 2002-01-01, #555>
Metadata anchor
(MD_ID*)
<#555>
Source system attribute
(MD_ID*, MD_SRC_Metadata_SourceSystem)
<#555, 'A'>
Recording time attribute
(MD_ID*, MD_REC_Metadata_RecordingTime)
<#555, 2001-08-20>
The connection between the models is on PR_PRI_Metadata = MD_ID. Primary keys indicated by *.
I hope that clarifies how we use recording time in mono-temporal AM. Note that I am working on a bi-temporal extension, in which recording time if lifted from the metadata model to the data model. I expect it to be ready by the end of the year.