diff --git a/net-doc/net/scribblings/net.scrbl b/net-doc/net/scribblings/net.scrbl
index 7f31fb982f..a289e4aa66 100644
--- a/net-doc/net/scribblings/net.scrbl
+++ b/net-doc/net/scribblings/net.scrbl
@@ -13,6 +13,7 @@
 @include-section["smtp.scrbl"]
 @include-section["sendmail.scrbl"]
 @include-section["head.scrbl"]
+@include-section["unihead.scrbl"]
 @include-section["imap.scrbl"]
 @include-section["pop3.scrbl"]
 @include-section["mime.scrbl"]
diff --git a/net-doc/net/scribblings/unihead.scrbl b/net-doc/net/scribblings/unihead.scrbl
new file mode 100644
index 0000000000..b26a1a7c3d
--- /dev/null
+++ b/net-doc/net/scribblings/unihead.scrbl
@@ -0,0 +1,63 @@
+#lang scribble/doc
+@(require "common.rkt"
+          scribble/eval
+          scribble/bnf
+          (for-label net/unihead))
+
+@(define unihead-eval (make-base-eval))
+@interaction-eval[#:eval unihead-eval (require net/unihead)]
+
+@(define l1-seq @elem{@litchar{=?ISO-8859-1?Q?}...@litchar{?=}})
+@(define uni-seq @elem{@litchar{=?UTF-8?B?}...@litchar{?=}})
+
+@(define enc-format @elem{@litchar{=?}@nonterm{encoding}@litchar{?}@nonterm{transport}@litchar{?}...@litchar{?=}})
+
+@title[#:tag "unihead"]{Header Field Encoding}
+
+@defmodule[net/unihead]{The @racketmodname[net/unihead] module
+provides utilities for encoding and decoding header fields using the
+@|enc-format| format.}
+
+@defproc[(encode-for-header [s string?]) string?]{
+
+Encodes @racket[s] for use in a header.
+
+If @racket[s] contains only ASCII characters, then the result string
+will have the same content as the given string. If @racket[s] contains
+only Latin-1 characters, then on each CRLF-delimited line, the
+space-delimited sequence containing all non-ASCII characters in
+@racket[s] is encoded with a @|l1-seq| sequence. If @racket[s]
+contains non-Latin-1 characters, then on each CRLF-delimited line, a
+space-delimited sequence containing all non-ASCII characters in
+@racket[s] is encoded with a @|uni-seq| sequence.
+
+@examples[
+#:eval unihead-eval
+(encode-for-header "English")
+(encode-for-header "français")
+(encode-for-header "→" )
+(encode-for-header "→\r\nboth → and français here")
+]}
+
+@defproc[(decode-for-header [s string?]) string?]{
+
+Decodes header fields that use the @|enc-format| encoding format. The
+specified @nonterm{encoding} is generalized via
+@racket[generalize-encoding] before decoding content.
+
+@examples[
+#:eval unihead-eval
+(decode-for-header "English")
+(decode-for-header "=?UTF-8?B?4oaS?= =?ISO-8859-1?Q?fran=E7ais?=")
+]}
+
+
+@defproc[(generalize-encoding [s (or string? bytes?)]) (or string? bytes?)]{
+
+Generalizes the encoding name @racket[s] to compensate for typical
+mailer bugs: Latin-1 and ASCII encodings are geenralized to
+WINDOWS-1252; GB and GB2312 are generalized to GBK; and KS_C_5601-1987
+is generalized to CP949.}
+
+
+@close-eval[unihead-eval]