2

I submitted a transcript for the summarizer to summarize, but it gets an error thinking that the input is a URL instead.
The parser is probably only looking for a colon in the beginning word, where instead it should probably be something like ^\s*https?:// to determine if it's a URL.

The summarize works when removing "JOHN:" from the beginning.

I expected the summarizer to summarize the transcript.

    __ changed the title to Summarizer does not summarize text if the first word has a colon. .

      The problem is that they still use the javascript URL parser for client side validation (to check if the input text is a URL or a text). Which is definitely not the right choice for this.

      Snippet from summarizer code:

      try {
              new URL(text_input)
              is_url = true
      } catch {}
        a month later
        No one is typing