Looks cool, well done. I toyed with the idea of building something similar as a product a few months ago. I think there's an opportunity for a service that allow users to create their own integrations for services that are too niche for Alexa/Siri etc to bother with, but would still benefit from automation. Example might be booking local tennis courts, seeing what specials are on at local restaurant, etc.
Some other ideas-
- Save macros retroactively. i.e, Extension records all the time, user selects previous n steps from 'action history' view.
- Paid service for running macros in cloud on schedule, receive reports, notifications, etc.
- Invoke macros from Alexa, FB bot, etc
- With above, can interactively ask for instruction if it encounters an error (button not found, etc).
Thank you grw_ :) And especially for sharing your well thought out points.
I have noted down all your ideas for further exploration. I'm chewing on an idea of crowdsourcing + marketplace, where people can make use of the flows made by others (without sensitive credentials), and where users with complex automation requests can have someone do it for a small goodwill fee.
It's for non-developers and business users, but your demo uses vi and a command line. I would not expect most non-developers and business users to be comfortable with those tools.
Thanks wffurr for your feedback, using notepad will do as well, but yes, I believe there is still a large potential gap to be closed in removing user friction, and that is the priority in pipeline. The other being porting over to Windows after the ground issues from incoming feedback are resolved.
For the beta, a primary aim is to minimize writing code to zero or minimal if a user knows JavaScript. I started working on this general tool 2-3 months ago because I hate writing code to drive automation.
I was spending much time copying and pasting and editing that I can't believe its 2016. I want to retain all the capability of handling complex scenarios, without writing large chucks of code.
Do you mean cross-platform desktop apps base on Electron for example? Currently the terminate points are PhantomJS or Firefox and I haven't tested that use case.
For raw desktop apps, might be able to work some integration out with projects such as RobotJS or pywinauto. Regardless, feel free to reach me at support@tebel.org or the repo issues page.
And for non-developers the screenshot-driven Kantu Web Automation Browser (Chromium-based, https://kantu.io ) is much easier to use than any tool that relies on Xpath and CSS Selectors.
I have not used iMacros before but did a quick check. I guess one of the difference would be it costs between $500-$3000 for different types of licenses (common for RPA software like Automation Anywhere or UiPath), while TA.Gui is under MIT license free to use/modify/share.
iMacros supports scripting by writing code, while one of the primary aims of TA.Gui is to minimize writing code while still supporting datatables, object repositories, incoming and outgoing API calls etc. I can't add more since I have not used iMacro but have noted this tool for reference.
For Kantu, it reminds me of SikuliX project, except that SikuliX is for entire desktop while Kantu is for browser. Have noted this down as well for reference :)
Recently i have started working with NightmareJs to scrape few sites. There also you have to mention the flow as in click here , grab this data n all. Wondering how is it different from that . Also can we use it to scrape data from a site and save to database?
Hi uberneo, thanks for your feedback. TA.Gui is based on CasperJS (with PhantomJS as underlying browser in invisible mode or SlimerJS+Firefox in visible mode).
One of the main aims of TA.Gui is converting almost natural language into working JavaScript code to run automation. Of course, it supports JavaScript directly within the automation flow for developers to still retain their full expressive ability.
I haven't used NightmareJS before to comment on the difference between it and CasperJS. But yes, TA.Gui can be used to scrape data from a site and outputs to text files or screen. Current beta version (v0.6.0) does not support writing directly to database, however a script (shell, PHP, etc) should be able to directly read the output text files into a database.
If you are already familiar with NightmareJS, using TA.Gui would probably be a breeze for you. You might have a try to see and I appreciate feedback on gaps to support@tebel.org or through the GitHub repo issues page :)
Windows defenders reports a malware when trying to download the master branch. Now I will be forced to format my company laptop. Great! P.S: The reported virus is Trojan:Win32/Sprinsky.
Why on earth would you format it if Windows Defender caught the "virus" (in this case a false positive) before you had a chance to run it? As long as you haven't run it (php in this case, so you'd have to have run them on a local web server), even if there were a virus, you could simply delete the file without opening/running it and you wouldn't have an issue.
I also have to know...if you think you have to format the laptop whenever a file you haven't run is detected to be a virus, how many hours have you spent doing this useless exercise?
Hi register, I'm sorry for the trouble of false-positive. I have closed the issue raised with the following remark. Thanks to tinodotim, downandout, swznd and all for helping to isolate the cause!
Resurrectio_LICENSE.GPL is cloned from Resurrectio project. I will temporarily remove this file and contact Eric Brehault. I would still like to use the original copy of his license instead of adding a new general GPLv2.
Eric has provided an updated license which is verified to yield no false-positive. Has committed and updated the issue with URL of the scan results - https://github.com/tebelorg/TA.Gui/issues/8
Thanks supremesaboteur for your feedback, I have not used Dejaclick before. For Selenium IDE, I got the impression that I cannot have access to conditions, loops, datatables, object repositories, invoking by API/scheduling, make outgoing API calls, etc. Without writing code.
I thought it was a rather limited tool (the Selenium IDE part), so usually test automation teams write their own frameworks in their preferred programming language to drive Selenium WebDriver (which of course is mainstream), then write code to do the test automation.
I'm basically attempting to bring that sort of framework out of the picture, to let users being able to do the same kind of complex stuff without writing code. They can still use JavaScript code in their automation flow if they want.
Thanks LordKano, feel free to feedback any bugs or suggestions to support@tebel.org or through the GitHub repo issues page. I welcome suggestions to make it better!
Some other ideas-
- Save macros retroactively. i.e, Extension records all the time, user selects previous n steps from 'action history' view.
- Paid service for running macros in cloud on schedule, receive reports, notifications, etc.
- Invoke macros from Alexa, FB bot, etc
- With above, can interactively ask for instruction if it encounters an error (button not found, etc).