ThoughtTrail is a company with the aim of bringing us out of the infomation flood using flexible artifcical intelligence techniques.

ThoughtTrail unleashes information from disparate sources, globally reducing friction between walled gardens in Desktop software and Online web applications, offering strikingly practical building blocks for the Semantic Web, and Semantic Desktop to solve problems in corporate, non-profit and personal situations alike.

Instopix is a program for instant messenger users that automatically shows pictures relating to their conversations in a stream of visual imagery, inspired by the concepts of Visual Language.

Instopix generates revenue using related Amazon product referrals. We offer a new form of non-intrusive advertising which is content centric, adding non-intrusive commercial links only after the fact, if available, and if they are wanted, signified by the user clicking a given image from the Instopix stream where, only then is our referral link visible (as well as Wikipedia information on the selected topic). Instant Messenger users can give friends the download link to Instopix, taking advantage of the fastest viral marketing ever.

Instopix is part of a broader plan for producing a symbiotic economy of semantic desktop and semantic web solutions, providing a more comprehensive way for humans to express themselves by taking advantage of the emotional and cognitive cues provided by visual language.

Instopix is built upon, what we're calling: A Semantic Desktop Metadata Framework and Rapid Inference Engine Toolkit. Our flexible platform allows us to take text input from any instant messenger, web browser, or other application in order to infer what the user is interested in, based on concrete textual data.

Instopix currently uses Yahoo's Contextual API to identify phrases in peoples conversations. but because of ThoughtTrail's rapid fact aggregation technology, we can do clever things like:

• Using Google Suggest data to access a ranked statistical array of phrases in both the popular, and niche zeitgeist

• See if the phrases match an identifiable topic area using Bayesian inference, clustering and similarity modeling

• Easily switch between Google, Yahoo, Flickr, or user submitted Images

• Rank images and topics which specific types of users are interested in

Using collaborative filtering, we aim to understand peoples interests and predict what topics Instopix should show, as well as letting other programs and services take advantage of these insights, contributing to a generic platform of tools, to bring people out on top of the flood of information with the ThoughtTrail Platform.

Instopix can be thought of as a chess piece, in a long term strategy to model all the users interests, store all the users important files, documents, settings, intentions, events and factoids no-matter where on the internet or computer they exist. What's more, is that it can all be done in an open and transparent way.

Check out the mock-up from the pitch video:



Here is an actual screenshot of the alpha we're working on:



Conceptual Diagram for ThoughtTrail



There is more information on the Smart Metadata Framework on: http://thoughttrail.com/preview/