In a previous post I have uploaded an updated Chatscript front-end that uses Google Chrome’s Speech API to Github. 
https://www.chatbots.org/ai_zone/viewthread/2173/

I also added a version with a picture of an avatar.  The picture of the avatar can be changed in Chatscript by sending an embedded command in your chatbot’s text response.  You can specify the image name directly or specify an emotion that is mapped to an image file in an output macro.  So if you supply an avatar image for each emotion then you can change the expression on your avatar as you chat.

Now, I have experimented with “actions” so instead of specifying an avatar image, ie [avatar=happy.jpg] I can send a command [action=smile] and the front end contains code that executes when it sees the command smile.  The code could be anything but I have set it to show the happy.jpg for a short time and then revert back to the previous avatar image.  So far I have similar actions for wink, cry, and frown. 

I also added a “talk animation” which basically switches an image from “mouth closed” to “mouth open” and alternates while the TTS is active.  It is primitive but it is a start.

I have a working personal version using avatar art from http://www.pickaface.net.  After rereading the licensing I cannot repost the “free avatar” art online.

If there is anyone who would like to collaborate on a 2D “animated” web front-end for Chatscript then let me know.  I would like 100% free avatar art of “medium to high” quality that shows emotions plus “talking”.  I had created 12 static images on pickaface.net that are ~200px square.  They consist of the same forward facing avatar image with just the mouth and eyebrows modified.  I would prefer art and not photos and would like to keep it simple enough for others to be able to modify if needed without requiring Unity or a degree in animation.

In the meantime, I may try my hand at some simple avatar drawings myself.

Once I have something decent looking I will upload it to github.

I would like to implement blinking, eye movement, and small face movement next.

If there is no chance you can contribute artwork but are interested in an animated web-based front end for Chatscript and have some ideas then please leave your comments.

Would one of the avatars on this page be a suitable starting point?

http://opengameart.org/content/avatar-generator-with-15-trillion-combinations

These are available under a CC-BY licence.

Alternatively, this is a simple js library which appears to offer some of what you need.  The demo is here: http://dumbmatter.com/facesjs/
and the code is on github under an MIT licence.

Hi Alaric.  Doug Danforth and I have been working with animated characters that connect to ChatScript for a while now with our Virtual Patient Project (http://www.kellenmaicher.com/the-virtual-patient.html)

We use Unity and 3D characters I designed starting with base templates from the Autodesk Character Generator.  The characters are capable of emotional expression and blending based on responses text responses from ChatScript. Like your system, we send back actions from ChatScript with the responses that tell Unity how to blend the character emotions and behaviors.

The project began as a standalone app and morphed into a web-based app. We were creating Unity builds that would communicate with ChatScript using the Unity Web Player export format.

Recently, we were working on addressing the problem of Chrome deprecating NPAPI plugins, which the Unity web player relies upon. In September, Unity web player builds will no longer function in Chrome, a fairly serious problem since 65% of users use Chrome.

We seem to have resolved this issue though using helpful advice from this forum and by implementing websockify, a web sockets proxy that now allows us to create WebGL builds of our Virtual Patients.

Speech for our web builds has been on our “To-Do list” for a while now.  Since our original standalone solution utilized the Windows Speech API, we were unable to implement speech for our web builds, instead just relying on text.  Your solution seems to offer us a new possibility for integrating speech in our Unity apps that I’m extremely interested in exploring.

Regarding your own experiments, if you would like I can take screen caps of one of our characters with whatever expressions you may need.  This would not be very diffficult for me at all.

However, the characters are of course 3D…and you mentioned in your post you were looking for 2D artwork that could be easily modified.  If you have been unable to find a free online generator that will meet your needs, I might be able to create a character in an Illustrator or Photoshop file that you could quickly modify to create new expressions. Depending on your timetable, 2D animations might also be an option.