A new voluntary code means chatroom providers will have to offer safety features to protect children from online paedophiles.

The guidelines, Models of Good Practice, were drawn up by the Home Office with the help of the industry itself.

Most UK-based chatrooms should now provide safety tools in several different areas:

• The chatroom introduction
  There should be "clear, prominent" information about the kind of service offered, and the audience at which it is aimed.

  Whether the chatroom is moderated or unmoderated should be made clear.

• Safety advice
  There should be "clear, prominent and accessible" safety messages on front pages and in chatrooms themselves.

  There should be clear safety advice

  Links should be available to online safety guides, either internal or external sites such as www.chatdanger.com.

• Safety tools
  Chatrooms should include some or all of the following:

  Ignore buttons; alert buttons; "grab and print" and reporting mechanisms.

  Advice on handling abusive chatters.

  A filtering mechanism to pick up bad language, or prevent children giving out their e-mail addresses.

  A way for the user to block private chat or instant messaging.

• Moderated chat
  There should be an alert system, like a panic button, at the top of each page.

  All chat moderators should be properly recruited, screened, trained and supervised.

  Users should be able to complain about any problems they find with the moderator.

• Registration
  Personal information gathered at registration should be as limited as possible.

  The reasons for needing the information gathered, and how it may be used, should be clearly and prominently explained.

• Public profile pages
  In these pages - where surfers give information about themselves, including contact details, if desired - the user should be able to limit what personal information is made public.

  Children should be warned to be cautious.

  Children should be encouraged not to post their phone numbers, addresses, or e-mail addresses.

• Responding to problems
  There should be a clear system of receiving and responding to reports of incidents.

  Users should be told how long they will have to wait for a response, and what will happen.