Single page search mashup with chat
New here? Learn about Bountify and follow @bountify to get notified of new bounties! x

Basically, taking this and using a realtime / js framework, and adding realtime chat with other searchers.


  1. Allow users to enter a search term, use the google auto search to suggest:

    1. Return results from using Google Search JS API (even though it's deprecated).
  2. Return top 10 google results in a column on the left hand side, showing url, a bit of text

  3. For the first search URL returned by Google, pass URL to

  • must use the mightbuy scraper api, with a callback
  • mightbuy scraper will return a image list, and price, you will format and display it

it will look somethinglike this:

  1. Each of the google results in the left hand can be clicked on to see the results in the middle

  2. If the "Save" button for the middle result is clicked, it moves over to the search results over to the right, list of “saved” results

  3. Winning solution - add chat to this system!

You have to figure out how. Something like:
-- see others search terms, anybody on these site, who saves a result.
-- you can click on a saved search term, add a comment
-- You can see if others add a comment on your saved search term

User story: I search “ford f150”. Results are auto suggested to me (like they are on A list of 10 Google results appears on the left side, almost immediately. These are the exact same links that would be returned from google (using the api).

The first returned google url is passed to scraper, so I begin seeing an image loading and price showing (takes time). I then click on the 3rd google result in the left hand side, which updates the picture and price in the middle section. Any google result I click on the left side shows in the middle.

I then “save” that result, so it's seen on the right hand side, in Saved results. At this point I don't create an account (this is must a sample app), but is saved in session (and perhaps when I come back?)

In the chat results window, my saved search result is visible to other users on the site. If they click on the saved items they see scrolling by, they can leave a comment, which pops up on my saved item window.

So.. not a lot of details on chat.. you, the winning coder, should create a solution here!

This seems like a pretty huge task for the price.
Blender over 8 years ago
I agree with Blender.
noah-freitas over 8 years ago
yea.. I think those are fair comments. If I get no bids I'll probably do it myself, good excuse to try a new framework, I especially like ember.js.
Choppen5 over 8 years ago
I agree with the first comment
methanol over 8 years ago
If you took out the chat part (or just required a chat interface in the JS front-end with a predefined server API [not implemented for bounty]) you might have more luck.
noah-freitas over 8 years ago
@Choppen5 About the Google autocomplete, its basically impossible for a JavaScript-only retrieval of the file because same origin policy will not allow the JS to access the data inside received by the browser.
alex over 8 years ago
@alex The autocomplete feature seems to be implemented on Couldn't find much documentation for the google API, but the source of that page might help. Also found this SO post:
bevan over 8 years ago
@bevan, Thanks! I dug around and finally got it working.
alex over 8 years ago
@Choppen5, I think I have a decent 'chat'-like system... but instead of comments, it relies on upvotes, is that okay? And the upvote system uses some backend and I would like to know if you had some specific language you wanted me to code it in. Thanks!
alex over 8 years ago
@alex, that sounds great.. I could see that working, in the "chat" window you see the most upvoted searchs.. and if you "save" a search item you get a little buzz when somebody upvotes. Re:tech.. at this point, anything that works, no requirement as we are just prototyping.
Choppen5 over 8 years ago
awarded to alex

Crowdsource coding tasks.

2 Solutions

Winning solution

A demo is available here and you can download the files here.

How to use

  1. Create a new database / database user
  2. Edit manager.php with credentials for database
  3. Run the commands in the db.sql file in phpMyAdmin (I used PHP because that's the language that I'm most comfertable with)


  1. I changed some syntax around so the autocomplete and the results work basically immediately.
  2. Implemented "chat"
  3. Fixed a small bug with the handling of the item IDs


I used Bootstrap. I wrote the "chat" in PHP. The "chat" system updates every 2 seconds or when changed (upvote/new item added). People can upvote more than once and you can't delete posts over the UI (because there is no way to authenticate people here).

cool! For me autocomplete almost never worked, except when I back space, a few times. Sometimes the search results showed on the left as I was typing, and sometimes not. Any thoughts on chat?
Choppen5 over 8 years ago
@Choppen5 I'm trying to see if I can do anything about the autocomplete, and the search results are dependent on the google server. And for the chat, I'm still looking for a solution to that.
alex over 8 years ago
Well, great progress. If I can find anything to help with autocomplete from mightbuy/search I will send it.
Choppen5 over 8 years ago
@Choppen5, I just fixed that. Please try it again.
alex over 8 years ago
@tjerkw @sergiotapia The comments section is reserved for comments on this solution. I'll make that more clear in the next iteration. BTW, most users provide solutions on Bountify for fun, like on StackOverflow or, with bounty money being a nice bonus. Thanks for your understanding.
bevan over 8 years ago
Thanks Alex! Awarded the bounty.
Choppen5 over 8 years ago

Had the time to make a Rails + websockets solution with chat that uses most of @alex's front-end (he did all the real work, so don't accept my solution please).

Demo is here:

I'll post the repo soon, unfortunately I'm using a Pusher account with it and the keys are still in the code, so I don't want to expose that yet (I use that Pusher account elsewhere too).

View Timeline