html2txt is a command line program designed to convert from html to txt.
Syntax is:

html2txt [options] htmlfile [htmlfile] [htmlfile] . . .

If no html file name is provided, the program will read from stdin 
(usually the keyboard) until an end of file mark is encountered (^Z for 
DOS/Windows, ^D for *nix, and I haven't a clue for the Macintosh).

Text appearing in <h> tags is centered to a line length of 80 characters
by adding spaces to the beginning of the line.  If the line is longer than
the centering length, it will not be centered.  The -head option can be 
used to add additional indentation to header text.

Text appearing in <p> tags is output as paragraphs, appropriately indented.
Text within paragraphs is output as a single line, followed by a single
newline or CRLF pair, depending on OS.  Word wrapping is not performed.

By default, output is single spaced between paragraphs, paragraphs are 
indented by 5 spaces, the underscore character is used to indicate 
italicised or emphasized text, the asterisk character is used to indicate
bold or strongly emphasized text, and headers are centered to a line
length of 80 characters.

options include the following:

	-i word or
	-italic <word>  use <word> as the italic indicator.  The
					word 'none' will suppress italic presentation
    -b <word> or
	-bold <word>	use <word> as the bold-face indicator.  The
					word 'none' will suppress bold presentation
	-t <n> or
	-tab <n>		use <n> number of spaces to indent paragraphs
	-l <n> or
	-line <n>		set the line length to <n> for centering
	-s <n> or
	-spacing <n>	put <n-1> lines of blank text between paragraphs
	-v or
	-verbose		output warning messages about badly-formed html
	-f <filename> or
	-file <filename>	Write error and warning message to the 
					specified file rather than stderr
    -raw            leave chars > 128 unchanged upon output
    -ascii          use ASCII for output, Latin-1 for input
    -latin1         use Latin-1 for both input and output
    -iso2022        use ISO2022 for both input and output
    -utf8           use UTF-8 for both input and output
    -mac            use the Apple MacRoman character set

Because it reads from stdin, html2txt can be used to convert html which is
being output from another application.  For example, to use the unix
stream editor, sed, to replace the <br> tag with ending and beginning 
paragraph tags, then convert the resulting file to text you could type:

sed 's/<br>/</p><p>' bad.html | html2txt > good.txt

To turn off centering, set the line length to 0:

html2txt -l 0 in.html > out.txt

Another example:

html2txt -l 0 -h 15 in.html > out.txt

This would uncenter headings, and indent them by 15 spaces.  To remove 
almost all markup, use the following:

html2txt -i none -b none -t 0 -l 0 in.html > out.txt

This program was recently modified to accept, without complaint, Microsoft
pseudo-html as produced by Microsoft Word 2000.

html2.txt is available from http://www.dysfunctionals.org/~networker/html2txt.zip.