Edbrowse, a Command Line Editor Browser

From the home page and user's guide:

> Edbrowse is a combination editor, browser, and mail client that is 100% text based... The inspiration for this program came from Karl's many years of experience as a blind computer user.

> You can edit a URL as easily as a local file, and activate browse mode to...discard most of the formatting information and retain the links and fill-out forms. This allows blind users to access the Internet through an application that is compatible with the linear nature of speech or braille.

> The interface is similar to /bin/ed, though there are many more features, such as editing multiple files simultaneously, and rendering html.

> This program was originally written for blind users, but many sighted users have taken advantage of the unique scripting capabilities of this program, which can be found nowhere else. A batch job, or cron job, can access web pages on the internet, submit forms, and send email, with no human intervention whatsoever.

Thanks for posting this. I've actually seen it many years ago but never tried it. I'm not blind but I'm actually interesting in making UIs for blind people and the CLI seems to look perfect for this. Can you confirm that?

BTW, do you know if TUI programs (using curses/ncurses or related) are usable for blind users?

I just wrote a plugin to make Edbrowse work with gemini. It supports TOFU, prompts, client certificates, handles redirects, and passes most of the gemini torture tests, that don't involve crazy unicode wrapping. Edbrowse's plugin support is extremely basic, so telling it about other media types, and auto redirecting away from gemini isn't possible without modifications to Edbrowse, as far as I know. I plan to add a few more features, like the ability to view the server certificate, and move up a directory, but I've only put about six or so hours into this thing so far, having never even used a gemini client beforehand.

Here's the plugin portion that goes into your Edbrowse configuration (~/.ebrc; you'll need to update the paths for your local system). If you want to use a client certificate, add -c your-cafile.crt -C your-cakey.crt to the program = ... line, before %i.

    plugin {
        type = */*
        desc = Gemini
        protocol = gemini
        program = /Users/ncarpenter/.config/edbrowse/plugins/ebgmni.py %i
        outtype = h
    }

And here's the helper Python script. It's a simple gemini client that converts text/gemini responses to HTML, for display in Edbrowse. You'll need to pip install ignition-gemini. The above plugin's program = line should point to it.

https://gist.github.com/n0ot/b257e44a414732ef97ca08c00b5a7a9...

I use Edbrowse all the time, including to read Hacker News. You can add custom functions into ~/.ebrc, and I have one to search for the next post on an HN page, or the "More" link, if there are no more posts on the current page.

    function:hnn {
        /^\({} \|{More}\)
    }

Edbrowse supports JavaScript, but it doesn't work very well, so I turn it off at startup, and only turn it on if I think it might make a page work better (it usually doesn't). Here's a bit of my config:

    # I use certifi to provide the trusted root certificate authorities
    certfile=/usr/local/lib/python3.8/site-packages/certifi/cacert.pem
    jar=/Users/ncarpenter/.config/edbrowse/cookiejar
    downdir = /Users/ncarpenter/Downloads
    # Disable cache, so each Edbrowse session doesn't cache where I've been to disk.
    # It doesn't have persistent history anyways.
    cachesize=0
    
    # User agents (type ua0 for edbrowse, ua1 for Lynx, etc)
    agent = Lynx/2.8.4rel.1 libwww-FM/2.14
    agent = Mozilla/4.0 (compatible; MSIE 5.5; Windows 98; Win 9x 4.90)
    agent = Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 6.1; WOW64; Trident/5.0; SLCC2; .NET CLR 2.0.50727; .NET CLR 3.5.30729; .NET CLR 3.0.30729; Media Center PC 6.0; .NET4.0C; .NET4.0E)
    agent=Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/72.0.3112.113 Safari/537.36
    
    # Automatically convert Markdown, Epub, PDF to HTML
    plugin {
        type = markdown
        desc = markdown file
        suffix = md
        program = pandoc --toc -f markdown %i
        outtype = H
        down_url
    }
    
    plugin {
        type = epub
        desc = epub file
        suffix = epub
        content = application/epub+zip
        #  %o is the temp output file generated by the program
        program = pandoc -f epub %i
        outtype = H
        down_url
    }
    
    plugin {
        type = pdf
        desc = pdf file
        suffix = pdf
        content = application/pdf
        #  file must be local
        down_url
        program = pdftohtml -i -noframes %i %o >/dev/null
        outtype = H
    }
    
    # Play links to audio with mpv
    plugin {
        type = audio
        desc = streaming audio
        protocol = rtsp,pnm,sdp,pls
        suffix = rm,ra,ram,ogg,mp3,mp4,m3u,m3u8,opus,flac
        content = audio/x-scpls,audio/mpeg,application/pls+xml
        program = mpv --really-quiet --no-audio-display --no-ytdl --af=scaletempo=stride=20:overlap=1 %i
    }
    
    # Use mps-youtube to handle Youtube URLs
    plugin {
        type = audio/youtube
        desc = streaming audio from Youtube
        suffix=mpeg # Doesn't matter, but needed for edbrowse not to complain
        urlmatch = .youtube.com/watch?
        program = mpsyt url %i
    }
    
    # These commands run when Edbrowse starts
    # See the user's guide for what these commands mean, but I'm basically setting up startup preferences
    function:init {
        db0
        js-
        sg+
        ci+
        dx
        endm+
        rl+
        z37
        db1
    }
    
    function:b {
        b ~/.config/edbrowse/bookmarks
    }
    
    function+atb {
    db0
    A
        ,j
        w+ ~/.config/edbrowse/bookmarks
        ^
        db1
    }
    
    # Typing <hs some.domain.tld is easier than typing
    # b https://some.domain.tld
    function:hs {
        b https://~0
    }
    
    # Because I read RFCs enough to make use of this
    function+rfc {
        b https://tools.ietf.org/html/rfc~1
    }
    
    # Open the current page in the default browser,
    # when it just isn't working right in Edbrowse.
    function+o {
        db0
        ub
        !open "'_"
        b
        db1
    }
    
    # Copy the current URL to the clipboard
    function:cpp {
        !printf '%s' "'_" | sed ' s`^\([a-zA-Z0-9]\{1,\}://.*\)\(\.browse\)$`\1`' | tr -d '\n' | pbcopy
    }

I left a few functions out, like ones to search Google and DuckDuckGo, because I store the search forms locally, but can include those in a separate comment, if anyone's interested.

any screenshots/video?