The Semantic Engine

I’ve just came across with an open source program called “The Semantic Engine” that uses Latent semantic analysis (see
It is nowhere near DT but may be useful for some specific purposes. Currently Mac only but both Win and Lin versions are on the way.
It can index txt, rtf, pdf, doc and html. You can search, get related words to your word or related documents for a selected one. One unique feature is, after a semantic search you can see related words in a given number of clusterings you choose and see them in a nice visualization window.
It has a simple interface, you can not browse or edit your documents, and currently very unstable. But it may be helpful for some of you, if you want to look in and tweak the code for your needs.

It seems that our famed Devonthinker Steven Johnson is using this to set up his own search engine. You can search and read his famous database right here:

I wonder if any of you will come up with some useful usage scenario.


Not sure if it’s a useful search engine. Try the experiment I just ran:

Search for “Darwin” 0 results
Search for “evolution” 753 results; the first one contains the string “Darwin”

You are right, it is an odd behavior, I think it doesn’t accept capital letters in search queries. Searching for “darwin” gave 329 results. Though I think the interesting feature is the visualization in the stand alone application (which is very crashy right now).