Understanding Searching with DevonAgent

Can someone please help me understand how to use DevonAgent? For an example, I was curious about hiking in the Everglades where there are now constrictors. I enjoy tracking and wildlife in general. I was curious to research situational awareness with constrictors, and possible defense against an attack. So I tried using DevonAgent and selected Web (Deep) for my search.

Under Search Sets/Plugins I have every box checked except Bing - all of them in essence. I assumed this would be slower, which It was, but that I should find some sort of result. I searched for “defense against constrictors” (without quotes).

I got no useful information in the results nor did the magic hat associate any useful information. I got results from acronym finders, dictionaries but no real useful data.

I placed my search criterion in quotes in this post but I did not put the three words in quotes as I entered the search. How would one use DevonAgent in this way?

I know people seem to like this program but so far, I am not getting much use out of It. Any help in clarifying how this should be working would be appreciated.


DEVONagent looks for items that meet the search criteria.

Are there pages out there about hikers attacked by boa constrictors or pythons? In the Everglades? Without checking, i’d tend to doubt that. Even if so, would such a page then go on to recommend defense against the snakes?

I’ve seen a number of TV shows about constrictor snakes in the wild, often about capture of these creatures by someone who jumps into a swamp and grabs a snake. The constrictor snakes are usually shy around humans.

There are pages about pythons in the Everglades, and if you do a search for such pages DEVONagent Pro will find them. The appropriate criteria would not include “defense”. You will be much more likely to find good results by a search using pythons AND everglades as the criteria, probably without use of the term constrictor.

It was just an example, I plan to do more serious research. :smiley: Although, I used to be fascinated by herpetology, for years one never gave constricting snakes in the USA a second thought. Now apparently It is worth a second thought. But, point taken, using Boolean and thinking about results.

I also tested making a new search set called “Science" which included science and blog results and google and one other search engine. Keeping all that other “stuff” out seemed to help as well. I was getting all sorts of hits because I was searching in all sorts of places - some of them not very appropriate for the results I was seeking.

I posted this so I could learn a little better how others are making good use to this app. Using good logic in search terms and not over complicating where I search seems to be part of It. I am sure I have more to learn though.


Biggest advantage that I get with DevonAgent is that it filters out all the “fake” Amazon pages, Ebay pages etc.

When you normally get to page 4 and onwards on Google or BING search the quality is very low but with DevonAgent it is a more even quality “further down” in my searches.

i’ve had the most success with single terms in my searches. i wouldn’t expect to get much from “defense against constrictors.” for one thing, “defense” is a pretty general term, and i kind of doubt defending against constrictors is a trending topic or anything. i guess people recommend that you don’t fall asleep next to one of these snakes? “against” is a preposition, so pretty useless in any search. in my opinion, “constrictor” is the key term here. i’d do an agent search, have everything saved automatically as a one-page pdf in devonthink, and then search within those results for pages that contain key words like everglades or defense. just an idea. let us know how it goes.

Interesting … Clearly I do not understand how this works yet. Apparently DevonAgents charm lays in It’s ability to ingest pages and re-sort them for better results. Similar to DTPO I suppose, It’s ideal use would be to build It’s own database of results and then, search that information.

I am still thinking of it as trying for a specific search result when needed rather than building a repository of search results to sort through over time - which is my understanding of your comments?


A neat feature of DEVONagent Pro is the See Also button. If you select a search result that is an interesting one, See Also will suggest other pages that may be contextually similar, as well as contextually similar documents in open DEVONthink databases.

This can be helpful in itself, and also as a tool to look at terms used that might be useful for refining search criteria.

I was only suggesting a possibility for your project. Personally, I am most excited about DA’s regularly scheduled crawling of the Web searching for items related to my research topic, the gathering of stuff it finds, and the archiving of everything into my DT database. Every morning I wake up and I have a few dozen pages about my research ready and waiting to be read in my inbox – I often sync the content to my iPad and have the reading materials with me all day.

That’s just my workflow, though. As Bill said, there are other ways to use DA as well. Some people prefer not to save the pages. They work with it in entirely different ways. Play around with the searching and see what works best for you. Check back and let us know. It’s always nice to hear new use cases.

This would be an interesting how-to to post here – if you have time – as a separate thread so as to not hijack the OP’s questions. These use cases are very helpful, I feel.

Another neat feature of scheduled searches by DEVONagent Pro: check the option to send an email listing the pages found in each run of the scheduled search. As I use Mail, the results are automatically available on my iPad or iPhone as clickable links to the results and I can review them in Safari on the iOS device.

That can be a productive way to while away time waiting for service on my car or attending a boring meeting. :slight_smile:

I have owned DAPro for years. I go back to it somewhat regularly. I’ve just never had success figuring it out to my benefit. Even following step by step. I am sure I am just not committing the time needed for the learning curve. Nothing against it really as I know many are happy with it, but it is just confusing to me. Maybe not setting up preferences right. Never figured out where the Archive button sent something… that still confuses me.

I’ve also battled, to be honest. I tried the trial a few times, but never really got to grips with the Search function, and abandoned it after a few hours each time.

That being said - I’m not criticising it, since I realise the ‘problem’ lies with me, in not spending enough time getting to learn how it works - but I don’t find it to be as easy to pick up and use.

No doubt when I’ve finished my current project, and want to start filtering more general web-based searches, I will make a more concerted effort. I suspect that, like with DTPO, I first need to get greater clarity on what it is that I want to do with it.

I suspect It has to do with the idea that DA is probably more useful to a person who knows exactly what they are looking for before they search. Scientific research, legal cases and the like. DA appears to be useful for gathering and helping to clarify results - when you have something in mind. That is that only reason I can think of as to why having DA continually refine searches would make any sense.

I learned long ago at a professional speakers meeting more about Google than I had previously understood. Google can do advanced searches using tools most of us don’t know much about, using boolean operators and the like. I also found out that there are paid access databases to search in like Lexus Nexus and others that can yield much more information than Google can. Those paid houses will also do research for you - for a fee of course. I am guessing that DA is a database research tool for everyman sort of thing. It is designed to build repositories of search results data of your own. Therefore It Isn’t about searching per se. Searching is part of what DA does but DA is really about building your own research database. That is the only reason I can think of why having It do regular searches and import results makes any sense at all.

I suppose DA is a useful tool if you are a very focused sort of individual or at least, you go after focused and pointed research topics most of the time.

What DA does NOT appear to be is a truly useful replacement for advanced Google searching. I think of searching usually in the moment. When I am fleshing out an idea, when I am writing about something, I want to know more about It. I Want immediate results. I want to know all about a person place or thing. I am less interested in tracking my results with repeated searches over time. I cannot even conceive of the usefulness of doing that. But a scientist would who was researching something. He or she would want to know how that topic was added to or clarified over time. A lawyer would want to. He or she would want to stay a breast with any information related to a case they were working on which could go on for years.

As the OP, my original thought was about looking for information about constrictors. If I had focused on a specific type of constrictor or a specific habitat where that constrictor might occur then the search would have been much better. I suppose that was my fault, chalk It up to operator error. Perhaps one reason DA is so hard to grasp for me is because sometimes I have a hard time being specific about what I am looking for. The see also does help with that, a lot.

Just my opinions of course. I welcome seasoned veterans here in the forums to clarify.

Post Script: I can see DA being very useful to me if I was giving a presentation or a speech to a company. I would want to know what was trending within the company and specifically what was happening related to my presentation topic. I would want to know about key personnel changes in the industry related to my client and I would want to know - If the presentation was to take place in another city, what was currently trending in that city, are they in a record drought, is crime up or down, is the housing market changing? I can see how not only knowing these things, but tracking them over time would help keep me informed and make my presentation much better. As I am on my flight to my presentation I can check my iPad to be sure I am still up on current trends and if not, incorporate that information into my presentation.

Most of the searches I do with DEVONagent Pro are not scheduled searches, although I do have some of those for keeping up with new information about a topic as it appears on the Web.

Many of my searches use a single term, especially when I’m initially looking for information. The more complex a search query, the more highly filtered the results. Adding any term that doesn’t exist in pages out in the wild, wild Web in conjunction with others will result in null results. Adding “defense” to your query about constrictors is an example.

I find DEVONagent more useful than Google advanced searches when I’m looking for new content for DEVONthink. The search results have already been dowloaded and can be easily added to DEVONthink.

Often, the Digest provided by DEVONagent is useful. For example, I needed to pull information about hexachlorobenzene and wanted to keep references in DEVONthink. I ended up saving the Digest to DEVONthink. It was almost a hundred pages of summaries of reference sources, with links. It was a great source of information about historical uses of hexachlorobenezene in consumer products, toxicology and case histories of pollution incidents.